framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,16384,26.751805623372395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,10240,16.61191134982639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,12288,17.026636759440105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,12288,20.152225070529514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,16384,22.71289571126302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,10240,14.979782104492188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,8192,9.467442830403646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,8192,21.941670735677082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,7168,8.978869120279947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,7168,11.41722191704644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,6144,6.921321868896484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,6144,9.950057135687933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,5120,8.06374782986111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,5120,5.584315405951606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,16384,13.934951782226562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,4096,6.306585947672526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,4096,3.8441378275553384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,12288,9.658490498860678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,10240,7.96415032280816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,3584,3.391447067260742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,3584,5.553830040825738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,3072,4.774376763237847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,3072,3.330890655517578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,7168,5.675575256347656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,8192,6.623835669623481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,2560,2.290651533338759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,2560,3.886665768093533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,2048,3.4667920006646047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,2048,1.797769758436415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,6144,5.038608974880643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,1536,3.537345674302843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,5120,4.1927642822265625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,1536,1.930367151896159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,1024,2.5893190171983504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,1024,1.5068978203667536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,4096,3.2783724466959634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,768,1.3148347006903756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,768,0.7971822420756022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,512,0.5684782134162055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,512,0.9370702107747396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,512,0.6530382368299696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,3584,2.7942807939317493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,256,0.587724420759413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,256,0.49664889441596133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,2560,2.141311009724935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,128,0.4821155336168077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,128,0.42963822682698566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,64,0.41181688838534886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,768,0.7750728925069174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,64,0.43802666664123535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,65536,32,0.41685599750942653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,65536,32,0.4541164504157172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,3072,2.469548543294271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,1536,1.2557368808322482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,65536,26.338282267252605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,16384,6.415952046712239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,2048,1.6975377400716145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,16384,5.161083645290799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,65536,24.503565470377605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,12288,3.8706203036838107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,12288,5.017035590277778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,10240,4.08540513780382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,10240,3.0979173448350696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,1024,0.9395537906222873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,8192,3.1788221995035806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,8192,2.1646870507134333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,128,0.44466045167711044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,65536,256,0.47335200839572483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,7168,2.743322584364149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,7168,1.7738123999701605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,6144,2.3887848324245877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,6144,1.55603822072347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,16384,3.1968258751763234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,5120,2.0537545945909286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,5120,1.2437822553846571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,65536,31.087941487630207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,12288,2.4349305894639754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,4096,0.9711120393541125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,4096,4.152097066243489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,10240,1.9684933556450739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,3584,0.8189573287963867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,3584,3.6498497856987844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,3072,3.159272935655382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,8192,1.669240845574273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,3072,1.5932640499538844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,2560,2.6338621775309243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,2560,1.3319252861870658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,7168,1.4812951617770727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,2048,0.7676551077100965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,2048,0.44730668597751194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,6144,1.19559997982449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,1536,0.6044480005900065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,1536,0.35244888729519314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,5120,0.9827466540866427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,1024,0.40788178973727757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,1024,0.2563982274797228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,3584,0.6908035808139377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,768,0.33016443252563477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,768,0.21267733309004042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,4096,0.775677310095893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,2560,0.5241271124945747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,512,0.23755288124084473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,512,0.17634577221340605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,256,0.1296586725446913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,256,0.12356089221106635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,3072,0.6090266439649794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,128,0.11414666970570882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,128,0.1138346725040012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,64,0.10584177573521932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,64,0.11985333760579427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,16384,32,0.10397955444124009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,16384,32,0.12122666835784912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,1536,0.3321946726904975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,2048,0.43069156010945636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,65536,19.956654866536457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,65536,21.183285183376736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,16384,12.629526774088541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,16384,7.133822123209636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,768,0.20842133627997506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,12288,3.560527801513672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,12288,2.726788414849175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,1024,0.24836799833509657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,10240,3.075828340318468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,10240,2.214196523030599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,128,0.11391377449035645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,8192,2.19160885281033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,256,0.12093599637349446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,8192,1.5618346532185872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,7168,1.9756516350640192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,7168,1.3613102171156142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,16384,512,0.16655377546946207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,6144,1.6769190894232855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,6144,1.154070218404134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,65536,10.185229831271702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,5120,1.676352924770779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,5120,1.2207679748535156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,16384,2.6015680101182723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,4096,1.1283564037746854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,4096,0.7502186563279895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,12288,1.9351209004720051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,3584,1.5065075556437175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,3584,0.6863475375705295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,10240,1.5098613103230794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,3072,0.9525261984931098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,3072,0.5303564601474339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,8192,1.271392928229438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,2560,0.692846245235867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,2560,0.45986753039889866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,7168,1.096334245469835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,2048,0.580810652838813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,2048,0.3464755482143826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,6144,0.8832400110032824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,1536,0.44912534289889866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,1536,0.2609208954705132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,5120,0.760501331753201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,1024,0.3008186552259657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,1024,0.19439289304945204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,3584,0.5280266867743598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,768,0.22780089908175996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,768,0.16742666562398276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,4096,0.6133804321289062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,512,0.17632177140977648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,512,0.1427288850148519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,3072,0.47017600801255965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,256,0.10699377457300822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,256,0.10886310868793064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,2560,0.38999022377861875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,128,0.08771733442942302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,128,0.09523200326495701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,64,0.08023377921846178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,64,0.09986578093634711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,12288,32,0.0794737802611457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,12288,32,0.09641422165764703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,768,0.15940799978044298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,65536,16.99912346733941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,1536,0.2568879922231038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,65536,15.357148912217882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,16384,3.929523468017578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,16384,3.0607742733425565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,2048,0.32151733504401314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,12288,3.0882169935438366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,12288,2.250766118367513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,1024,0.19010755750868055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,10240,2.2288568284776473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,10240,1.7196604410807292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,256,0.09408533573150635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,8192,1.9969893561469183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,8192,1.3217911190456815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,512,0.12880622016059026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,7168,1.6343119939168294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,12288,128,0.08794666661156549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,7168,1.1230604383680556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,6144,1.5283092922634547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,6144,0.9662159813774956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,65536,8.516055636935764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,5120,1.2893430921766493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,5120,0.9238284428914388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,16384,2.0190658569335938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,4096,1.0669938193427193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,4096,0.6186960008409288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,12288,1.5815529293484156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,3584,0.8939893510606555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,3584,0.5243262184990777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,10240,1.2318542268541124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,3072,0.6845359802246094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,3072,0.47432888878716367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,8192,1.0586239496866863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,2560,0.5941777759128147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,2560,0.3764968978034125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,7168,0.8975049124823676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,2048,0.49302313062879777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,2048,0.2973875469631619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,6144,0.7947111129760742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,1536,0.380603551864624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,5120,0.6312328974405924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,1536,0.23268267843458387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,1024,0.2958915498521593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,1024,0.1718133290608724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,3584,0.45235999425252277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,768,0.20530666245354545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,768,0.1402719948026869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,2560,0.3292542298634847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,512,0.15031554963853624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,512,0.11726489331987168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,4096,0.511341306898329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,256,0.1028711133533054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,256,0.08986400233374701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,3072,0.38373777601453996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,128,0.07630844248665704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,128,0.0746213330162896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,64,0.06951289044486152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,64,0.0782711108525594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,10240,32,0.06861155562930636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,10240,32,0.07942577865388659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,2048,0.2695111168755425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,65536,13.233210245768229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,65536,8.616867913140192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,768,0.13551466994815403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,16384,4.009039984809028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,16384,2.5507208506266275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,1536,0.2150275574790107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,12288,2.4059742821587458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,12288,1.502155515882704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,1024,0.16213156117333308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,10240,1.861819585164388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,10240,1.258376015557183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,128,0.07509599791632758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,256,0.07959555255042182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,8192,1.5664283964369032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,8192,0.945472929212782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,7168,1.4669058057996962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,7168,0.8715777926974826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,10240,512,0.1090915534231398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,6144,1.2029875649346244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,6144,0.745522658030192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,65536,6.91664801703559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,5120,0.9680631425645617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,5120,0.5988355742560493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,16384,1.6515110863579645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,4096,0.7432497872246636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,4096,0.46895021862453884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,12288,1.3584213256835938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,3584,0.8875466452704536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,3584,0.5233875380622016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,10240,1.0191368526882596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,3072,0.5845573213365343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,3072,0.32841422822740346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,8192,0.831318219502767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,2560,0.47778134875827366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,2560,0.2689635488722059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,7168,0.7133519914415148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,2048,0.38845867580837673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,2048,0.21809066666497123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,6144,0.6239377657572428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,1536,0.30842222107781303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,1536,0.16851377487182617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,5120,0.5160133573744032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,1024,0.21904444694519043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,1024,0.13436444600423178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,4096,0.4096666706932916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,768,0.1595804426405165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,768,0.10524800088670518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,3584,0.3615439997778998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,512,0.1127857764561971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,512,0.08590044577916463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,2560,0.26830845408969456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,256,0.06906488868925306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,256,0.06866933239830865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,3072,0.32331821653578013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,128,0.05953511264589098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,128,0.06577244732115003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,64,0.054570668273501925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,64,0.06472533278995089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,8192,32,0.05430044399367439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,8192,32,0.06482666730880737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,2048,0.22193066279093424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,65536,10.800198872884115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,65536,7.9466137356228295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,1536,0.1760035620795356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,16384,2.496270285712348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,768,0.11080177625020345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,16384,1.7465573416815863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,12288,2.0710506439208984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,12288,1.3389227125379775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,1024,0.13209777408176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,10240,1.6555858188205297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,10240,1.1282257503933377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,256,0.06340710984336005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,8192,1.3743599785698786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,8192,0.831982241736518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,512,0.08790222141477798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,7168,1.1886133617824979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,7168,0.7529137929280599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,8192,128,0.05996444490220812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,6144,1.0050346586439345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,6144,0.6500097910563151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,65536,11.993932088216146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,5120,2.1083200242784286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,5120,1.8155227237277562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,16384,1.4665786955091689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,4096,0.6477973196241591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,4096,0.41944177945454914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,12288,1.106419563293457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,3584,0.5932497978210449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,3584,0.3910186555650499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,10240,0.9075617260403104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,3072,0.4857938024732802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,3072,0.3257351186540392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,8192,0.7452693515353732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,2560,0.43844623035854763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,2560,0.2573484314812554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,7168,0.6341235372755263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,2048,0.35039912329779727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,2048,0.20348800553215873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,6144,0.5413635571797689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,1536,0.26109332508511013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,1536,0.15215288268195257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,5120,0.4539608955383301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,1024,0.1857422192891439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,1024,0.11894755893283421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,4096,0.3784257835812039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,768,0.13878844843970403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,768,0.10114755233128865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,3584,0.32548265986972386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,512,0.09906222422917683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,512,0.07734044392903645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,2560,0.24032534493340385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,256,0.06120622158050537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,256,0.06103822257783678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,3072,0.28497687975565594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,128,0.052552888790766396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,128,0.054634667105144925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,64,0.04792888959248861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,64,0.05617600017123752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,7168,32,0.04795022143257988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,7168,32,0.055516444974475436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,2048,0.1987555556827121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,65536,9.451197306315104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,65536,7.107654147677952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,1536,0.1571822166442871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,16384,2.1822382609049478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,16384,1.487862269083659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,768,0.09826844268374973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,12288,1.738999154832628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,12288,1.152264912923177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,1024,0.11821066008673774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,10240,1.4082098007202148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,10240,0.9944275750054253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,512,0.07793333133061726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,8192,1.1502560509575737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,8192,0.7708551088968912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,256,0.056998219754960805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,7168,0.9579893747965494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,7168,0.6785777939690484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,7168,128,0.05359733435842726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,6144,0.8631519741482205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,6144,0.5582070880466038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,65536,5.477443695068359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,5120,0.7359093560112847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,5120,0.459279113345676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,16384,1.2414941787719727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,4096,0.5728444523281521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,4096,0.38366132312350804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,12288,0.9584969414605035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,3584,0.5059537887573242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,3584,0.3108728991614448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,10240,0.776967101626926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,3072,0.4266142315334744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,3072,0.2634800010257297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,8192,0.6281733512878418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,2560,0.36960445510016543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,2560,0.22288534376356336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,7168,0.5687537723117405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,2048,0.3104417853885227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,2048,0.17707644568549263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,6144,0.47082842720879453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,1536,0.21638933817545572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,1536,0.12786577807532415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,5120,0.39563645256890195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,1024,0.1514328850640191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,1024,0.09461066457960342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,4096,0.3235208988189697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,768,0.12223822540707058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,768,0.08804266982608372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,3584,0.2824915515051948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,512,0.08597422308391994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,512,0.06823911269505818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,2560,0.20847288767496744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,256,0.05329066514968872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,256,0.05336888962321811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,3072,0.24225065443250868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,128,0.0463137792216407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,128,0.04741600155830383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,64,0.042113777663972646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,64,0.048398223188188344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,6144,32,0.04231199953291151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,6144,32,0.04863199922773573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,2048,0.17223733001285127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,65536,7.721611870659722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,65536,6.9344372219509545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,1536,0.13991288344065347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,16384,1.981853273179796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,16384,1.3100639979044597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,768,0.08660711182488336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,12288,1.3967591391669378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,12288,0.985504044426812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,1024,0.1038702196545071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,10240,1.1895252863566081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,10240,0.8119049072265625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,256,0.051334222157796226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,8192,0.8911128573947482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,8192,0.6172480053371853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,512,0.06924711333380805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,7168,0.7929777569240994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,7168,0.5589217609829372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,6144,128,0.047819554805755615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,6144,0.6819341977437338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,6144,0.4833395746019151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,65536,4.556766086154513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,5120,0.5971022182040745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,5120,0.391773329840766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,16384,1.1332729127671983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,4096,0.4591066572401259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,4096,0.3057520124647352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,12288,0.8211110962761773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,3584,0.41416088740030926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,3584,0.2916417651706272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,10240,0.6775084601508247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,3072,0.34859289063347715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,3072,0.2375875578986274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,8192,0.5443911022610134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,2560,0.2948257658216688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,2560,0.18950222598181832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,7168,0.46896711985270184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,2048,0.2356355455186632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,2048,0.14502488242255315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,6144,0.4075448777940538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,1536,0.1941839986377292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,1536,0.1165013313293457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,5120,0.33839554256863064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,1024,0.12824622790018717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,1024,0.08668088912963867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,4096,0.2718213399251302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,768,0.09881066613727146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,768,0.07032977872424655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,3584,0.24121689796447754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,512,0.07344444592793782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,512,0.058596445454491504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,2560,0.1805724435382419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,256,0.04955733484692044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,256,0.046126223272747464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,3072,0.21548710929022896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,128,0.03985155622164408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,128,0.04446666770511203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,64,0.035565333233939275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,64,0.04448089003562927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,5120,32,0.03636711173587375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,5120,32,0.04309689005215963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,2048,0.1497982210583157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,65536,6.051718393961589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,65536,4.4484782748752165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,1536,0.11918844117058648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,16384,1.497036404079861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,16384,1.0412604014078777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,768,0.07397600015004475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,12288,1.0697128507826064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,12288,0.8086471027798123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,1024,0.09043200148476495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,10240,0.9418711132473416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,10240,0.6546808878580729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,256,0.04362666606903076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,8192,0.7119270960489908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,8192,0.50527376598782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,512,0.05860266420576307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,7168,0.6309679879082574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,7168,0.4539635446336534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,5120,128,0.04062044289377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,6144,0.5469386842515733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,6144,0.38405688603719074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,65536,6.315318213568793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,5120,0.5418462223476833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,5120,0.32779733339945477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,16384,0.8752409087287055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,4096,0.3658124340905084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,4096,0.24728443887498644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,12288,0.655169751909044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,3584,0.3175537851121691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,3584,0.21506844626532662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,10240,0.5485528839959039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,3072,0.274146662818061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,3072,0.18728444311353895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,8192,0.4486417770385742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,2560,0.23300978872511122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,2560,0.15264444881015354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,7168,0.3845280011494954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,2048,0.18531555599636504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,2048,0.11842489242553711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,6144,0.3292880058288574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,1536,0.14115733570522732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,1536,0.09136355585522121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,5120,0.2799359957377116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,1024,0.10046222474839951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,1024,0.06822311215930514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,3584,0.20071378019120958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,768,0.07923999759886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,768,0.05796088774998983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,4096,0.2315297656589084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,512,0.06024710999594795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,512,0.04946310983763801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,2560,0.14883822864956325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,256,0.03789066606097751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,256,0.038945777548684016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,3072,0.17393244637383354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,128,0.03183199961980184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,128,0.03436622354719374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,64,0.02883111106024848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,64,0.035079999102486506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,4096,32,0.029734220769670274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,4096,32,0.03534133235613505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,2048,0.1254568894704183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,65536,5.327371385362413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,65536,3.863710191514757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,1536,0.09850755665037368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,16384,1.2857181760999892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,16384,0.9557128482394748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,768,0.062124444378746875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,12288,0.9970177544487847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,12288,0.7147662374708387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,1024,0.07412355475955539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,10240,0.788815975189209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,10240,0.6057599915398492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,256,0.0358497765329149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,8192,0.6436968909369575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,8192,0.46753512488471133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,512,0.049103998475604586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,7168,0.5498853259616429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,7168,0.4124435583750407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,4096,128,0.03341689043574863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,6144,0.4673573176066081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,6144,0.35469333330790204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,65536,3.4747066497802734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,5120,0.4002959993150499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,5120,0.2942062218983968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,16384,0.7947626643710666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,4096,0.3242257701026069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,4096,0.23307998975118002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,12288,0.599731551276313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,3584,0.29040445221794975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,3584,0.2043759955300225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,10240,0.5012746916876899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,3072,0.24232355753580728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,3072,0.17361777358584932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,8192,0.3920417891608344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,2560,0.20119910769992402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,2560,0.14551466041141087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,7168,0.3559528986612956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,2048,0.16857511467403838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,2048,0.11147467295328777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,6144,0.30010665787590873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,1536,0.12575822406344944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,1536,0.08337422211964925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,5120,0.2517208788130018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,1024,0.08862311310238308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,1024,0.06102399693595039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,4096,0.20484533574846056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,768,0.06992977857589722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,768,0.053415109713872276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,3584,0.18446489175160727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,512,0.05214933223194546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,512,0.044293334086736046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,2560,0.13618932829962835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,256,0.0343502230114407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,256,0.0358515547381507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,3072,0.15719732973310682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,128,0.02875111169285244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,128,0.03228711088498434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,64,0.02672622270054287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,64,0.032348444064458214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3584,32,0.02728355593151516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3584,32,0.032093332873450384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,2048,0.11250310473971897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,65536,4.708588494194879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,65536,3.3128053877088757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,1536,0.09068622191747029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,16384,1.1419137318929036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,16384,0.8528773519727918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,768,0.054752889606687755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,12288,0.8212240007188586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,12288,0.6368728743659126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,1024,0.06756177875730726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,10240,0.6693742010328504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,10240,0.533929771847195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,256,0.03224622209866842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,8192,0.537393781873915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,8192,0.4097910987006293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,512,0.0431182218922509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,7168,0.4724044269985623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,7168,0.36327555444505477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3584,128,0.02952444553375244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,6144,0.3995324505700006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,6144,0.31561067369249135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,65536,3.047580507066515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,5120,0.3642720116509332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,5120,0.26169244448343915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,16384,0.6972053315904406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,4096,0.2843688858879937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,4096,0.20455200142330596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,12288,0.5383395618862575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,3584,0.2451982233259413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,3584,0.1803679996066623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,10240,0.4397351211971707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,3072,0.2059404452641805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,3072,0.15437511603037515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,8192,0.35322043630811906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,2560,0.17316177156236437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,2560,0.12881510787540013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,7168,0.3111742337544759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,2048,0.14110222127702501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,2048,0.09875555833180745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,6144,0.2655751175350613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,1536,0.10808710919486152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,1536,0.07472978035608928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,5120,0.2247840033637153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,1024,0.07669511106279162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,1024,0.05433244506518046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,4096,0.18140355745951334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,768,0.060495111677381724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,768,0.04819555415047539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,3584,0.16057066122690836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,512,0.044991112417644925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,512,0.03919555412398444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,2560,0.12030577659606934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,256,0.030365334616767034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,256,0.031705776850382485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,3072,0.14270844724443224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,128,0.025763556361198425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,128,0.028612444798151653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,64,0.023355555203225877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,64,0.028386665715111628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,3072,32,0.02482844392458598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,3072,32,0.028538667493396338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,2048,0.1008311112721761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,65536,3.9902809990776906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,65536,2.864707522922092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,1536,0.08124799860848321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,16384,0.888494226667616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,16384,0.7163084348042806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,768,0.04915111263593038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,12288,0.687862237294515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,12288,0.5507582028706869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,1024,0.06031733089023166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,10240,0.5463013119167751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,10240,0.4693022304111057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,256,0.029151999288135107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,8192,0.4347626633114285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,8192,0.3621724446614583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,128,0.02602044410175747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,7168,0.387318213780721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,7168,0.3211288981967502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,3072,512,0.038211554288864136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,6144,0.3300480047861735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,6144,0.27861422962612575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,65536,2.804973390367296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,5120,0.4635964499579535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,5120,0.23648354742262098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,16384,0.6096160146925185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,4096,0.22871022754245332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,4096,0.18266488446129692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,12288,0.4601457913716634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,3584,0.20584355460272896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,3584,0.16010488404168025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,10240,0.38280177116394043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,3072,0.17808889018164742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,3072,0.1388746632470025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,8192,0.3066791163550483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,2560,0.1452666653527154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,2560,0.11612178219689263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,7168,0.26772978570726186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,2048,0.11759200361039902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,2048,0.08825155761506821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,6144,0.23240889443291557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,1536,0.09167466560999553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,1536,0.06608444452285767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,5120,0.19561333126491973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,1024,0.06459199719958834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,1024,0.04775822162628174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,4096,0.16217066182030573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,768,0.053209778335359365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,768,0.042859554290771484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,3584,0.14236622386508518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,512,0.03977244430118137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,512,0.03442577852143182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,2560,0.10658844312032063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,256,0.028010666370391846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,256,0.028062221076753404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,3072,0.12489510907067193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,128,0.021983999345037673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,128,0.025314667158656653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,64,0.01959733333852556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,64,0.02497244377930959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2560,32,0.020406223005718656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2560,32,0.02509511179394192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,2048,0.08871733480029637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,65536,3.240449693467882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,65536,2.458617740207248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,768,0.04275022281540764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,16384,0.7121058040195041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,16384,0.6574853261311849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,1536,0.07046844561894734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,12288,0.5223173565334743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,12288,0.4692871305677626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,1024,0.0525546669960022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,10240,0.44352266523573136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,10240,0.40083999103969997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,256,0.025438222620222304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,8192,0.3563377857208252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,8192,0.3095937834845649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,512,0.03343377841843499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,7168,0.32089243994818795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,7168,0.27798488405015737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2560,128,0.02251111136542426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,6144,0.26989155345492893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,6144,0.2355253431532118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,65536,2.1828507317437067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,5120,0.2438702318403456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,5120,0.196798218621148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,16384,0.5116497675577799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,4096,0.18247644106547037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,4096,0.1612888839509752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,12288,0.3839493327670627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,3584,0.16224977705213758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,3584,0.13715999656253391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,10240,0.31947022014194065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,3072,0.13836088445451525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,3072,0.1205493344200982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,8192,0.25893155733744305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,2560,0.11883644262949626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,2560,0.10274577803081936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,7168,0.22939909829033744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,2048,0.09671822521421646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,2048,0.0768631100654602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,6144,0.19940355088975695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,1536,0.07597688833872478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,1536,0.05726222197214762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,5120,0.16751821835835776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,1024,0.05307733350329929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,1024,0.0413680010371738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,4096,0.1377884414460924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,768,0.04267288910018074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,768,0.037031110790040754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,3584,0.12068443828158909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,512,0.03241777751180861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,512,0.028601778878106013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,2560,0.08957511186599731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,256,0.02239022155602773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,256,0.023584889041052923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,3072,0.104184885819753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,128,0.018045332696702745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,128,0.020962667134073045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,64,0.01495911180973053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,64,0.02124711043304867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,2048,32,0.015442666080262927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,2048,32,0.021866665946112737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,2048,0.07530844211578369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,65536,2.267213397555881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,65536,2.13682132297092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,1536,0.060876442326439746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,16384,0.5277742279900445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,16384,0.5281839900546604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,768,0.0369777778784434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,12288,0.3946559958987766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,12288,0.4038088851504856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,1024,0.04522933231459724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,10240,0.3372684319814046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,10240,0.33979111247592503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,256,0.02128177715672387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,8192,0.2637217839558919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,8192,0.2674230999416775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,512,0.028747555282380845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,7168,0.2339857684241401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,7168,0.23200623194376627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,2048,128,0.019440000255902607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,6144,0.19785600238376191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,6144,0.20075021849738228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,65536,1.7200613021850586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,5120,0.17900621891021729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,5120,0.1691875590218438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,16384,0.42279911041259766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,4096,0.13764799965752497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,4096,0.1369511153962877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,12288,0.3198808828989665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,3584,0.12139556143018936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,3584,0.12078400452931722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,10240,0.2680373456743029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,3072,0.10499378045399983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,3072,0.10355555348926121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,8192,0.21748444769117567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,2560,0.09108889102935791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,2560,0.0866231123606364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,7168,0.19007466899024117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,2048,0.07251911030875312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,2048,0.06561866733762953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,6144,0.16632266839345297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,1536,0.056735111607445605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,1536,0.049943112664752536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,5120,0.13887911372714573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,1024,0.040758222341537476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,1024,0.03554933269818624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,3584,0.10076622168223064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,768,0.03347288899951511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,768,0.03152622116936578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,4096,0.11323200331793891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,512,0.025644444757037695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,512,0.023817777633666992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,3072,0.08855199813842773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,256,0.018367111682891846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,256,0.019498666127522785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,2560,0.0768773357073466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,128,0.014084445105658637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,128,0.017272889614105225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,64,0.012067555553383298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,64,0.017664000391960144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1536,32,0.012681778106424542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1536,32,0.017821333474583097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,2048,0.06446488698323567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,65536,1.5294906828138564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,65536,1.72362306382921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,1536,0.051761776208877563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,16384,0.3549075656467014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,16384,0.4292151133219401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,768,0.0310808883772956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,12288,0.26289600796169704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,12288,0.32616088125440806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,1024,0.03836444351408217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,10240,0.22478045357598198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,10240,0.2747848828633626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,256,0.01856355534659492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,8192,0.17759199937184653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,8192,0.21680088837941489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,512,0.0247653325398763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,7168,0.15797421667310926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,7168,0.19041511747572157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1536,128,0.016410667035314772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,6144,0.13611822658114964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,6144,0.16532177395290798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,65536,1.373687956068251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,5120,0.11743733617994521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,5120,0.13996267318725586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,16384,0.3381022347344293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,4096,0.09534222549862331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,4096,0.11246221595340306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,12288,0.25334488021002877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,3584,0.08144978019926283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,3584,0.09931911362542047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,10240,0.21192533440060088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,3072,0.06981244352128771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,3072,0.08545777532789443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,8192,0.17388444476657441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,2560,0.06031555599636502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,2560,0.07136355506049262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,7168,0.1516017781363593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,2048,0.04881688952445984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,2048,0.05517333414819506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,6144,0.13134933842553034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,1536,0.03831733266512553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,1536,0.04095377855830722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,5120,0.11159200138515896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,1024,0.027279110418425664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,1024,0.029409776131312054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,4096,0.09202044539981419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,768,0.02196888956758711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,768,0.026722666290071275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,3584,0.08205866813659668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,512,0.01681155628628201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,512,0.019631110959582858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,2560,0.062056885825263135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,256,0.01202044470442666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,256,0.01607911123169793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,3072,0.07183911403020223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,128,0.009506666825877296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,128,0.01386755539311303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,64,0.007819555699825287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,64,0.013832888669437833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,1024,32,0.007986666427718269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,1024,32,0.01388533330625958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,2048,0.052686221069759794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,65536,1.0772595935397677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,65536,1.5107022391425238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,1536,0.042062222957611084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,16384,0.2683493296305339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,16384,0.3776293330722385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,768,0.02509511179394192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,12288,0.20415022638108996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,12288,0.28880977630615234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,1024,0.030729777283138696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,10240,0.1711395581563314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,10240,0.24214755164252388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,256,0.015248888068728976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,8192,0.14099733034769693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,8192,0.19152266449398467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,512,0.020383111304706998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,7168,0.12123200628492568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,7168,0.16847288608551025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,1024,128,0.01370844410525428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,6144,0.10295110940933228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,6144,0.14575910568237305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,65536,1.1824222140842013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,5120,0.08857866790559556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,5120,0.12295733557807074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,16384,0.29231378767225474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,4096,0.07164711422390409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,4096,0.0991520020696852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,12288,0.21909777323404947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,3584,0.062195552719963916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,3584,0.08822933170530532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,10240,0.1842648850546943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,3072,0.054543998506334096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,3072,0.07551999886830647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,8192,0.1484302282333374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,2560,0.04607999987072415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,2560,0.06334844562742445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,7168,0.13180088996887207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,2048,0.038514667087131076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,2048,0.0498577786816491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,6144,0.11414044433169895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,1536,0.029426667425367568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,1536,0.03661333190070258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,5120,0.09614844454659356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,1024,0.02091111077202691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,1024,0.025648888614442613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,4096,0.0792879992061191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,768,0.01718844473361969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,768,0.02400622268517812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,3584,0.07150044706132677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,512,0.013552889227867126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,512,0.01759733259677887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,2560,0.05388799972004361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,256,0.009568889107969072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,256,0.013998222019937305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,3072,0.06188978089226616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,128,0.007267555428875818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,128,0.01202933324707879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,64,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,64,0.012057777908113269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,768,32,0.006529777828190062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,768,32,0.012114666402339935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,2048,0.045520000987582736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,65536,0.7213546435038248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,65536,1.307334264119466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,768,0.02238222294383579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,16384,0.17812532848782012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,16384,0.3284613291422526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,1536,0.03647022114859687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,12288,0.14165866374969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,12288,0.25168000327216256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,1024,0.026271111435360376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,10240,0.11478843953874375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,10240,0.21010755168067086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,256,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,8192,0.09338044457965428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,8192,0.16644622219933405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,512,0.018053332964579265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,7168,0.08174400197135077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,7168,0.14624444643656412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,768,128,0.011747555600272285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,6144,0.07044800122578938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,6144,0.12705689006381565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,65536,0.9934053421020508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,5120,0.0608231094148424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,5120,0.10750222206115723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,16384,0.25015732977125377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,4096,0.04885422190030416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,4096,0.08666044473648071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,12288,0.1891626649432712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,3584,0.04201422135035197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,3584,0.07667733563317193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,10240,0.16007733345031738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,3072,0.03780355718400743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,3072,0.0663822227054172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,8192,0.12934933768378362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,2560,0.03178222311867608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,2560,0.055778668986426465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,7168,0.11497244569990371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,2048,0.026271111435360376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,2048,0.04243555665016174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,6144,0.09933777650197347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,1536,0.020421332783169217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,1536,0.030988444884618122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,5120,0.08420444197124904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,1024,0.015406222807036506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,1024,0.02194044490655263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,4096,0.06945155726538764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,768,0.012847999731699625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,768,0.021038222644064162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,3584,0.06207999918195936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,512,0.01037333326207267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,512,0.014874666929244995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,2560,0.0472302238146464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,256,0.0075599998235702515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,256,0.011986666255527072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,3072,0.05455199877421061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,128,0.00612088872326745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,128,0.010242666635248397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,64,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,64,0.010537777509954242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,512,32,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,512,32,0.010753778119881948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,2048,0.039825777212778725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,65536,0.4124142328898112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,65536,1.1023715337117512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,1536,0.031353778309292264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,16384,0.1043013334274292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,16384,0.2828755643632677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,768,0.019536000159051683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,12288,0.07965511083602905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,12288,0.21560177538130018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,1024,0.022872888379626807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,10240,0.06646755668852064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,10240,0.18186488416459826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,256,0.012671999633312225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,8192,0.054135110643174916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,8192,0.14650489224327937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,512,0.016136889656384785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,7168,0.04784977767202589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,7168,0.12899288866255018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,512,128,0.01071200023094813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,6144,0.04270755582385593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,6144,0.11197422610388862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,65536,0.8869626786973742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,5120,0.04081066780620151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,5120,0.09418221977021958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,16384,0.2242622110578749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,4096,0.02994399931695726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,4096,0.07591288619571261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,12288,0.16985155476464164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,3584,0.027190221680535212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,3584,0.06705244382222493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,10240,0.14166755146450466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,3072,0.023969777756267126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,3072,0.05744622151056925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,8192,0.11392177475823297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,2560,0.020624889267815482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,2560,0.047818667358822294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,7168,0.1006079978413052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,2048,0.016985777351591323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,2048,0.035531553957197405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,6144,0.08770933416154648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,1536,0.013341333303186627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,1536,0.02545333405335744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,5120,0.07337866889105903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,1024,0.009958222508430481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,1024,0.01850044396188524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,4096,0.06018133295906914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,768,0.008949333594905006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,768,0.018352889352374606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,3584,0.05330844389067756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,512,0.007227555745177799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,512,0.012795555922720166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,2560,0.040082666609022356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,256,0.005335999859703912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,256,0.010465777582592435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,3072,0.04616533385382759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,128,0.004389333228270213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,128,0.008782222039169734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,64,0.004088888979620404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,64,0.0090844440791342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,256,32,0.004292444636424382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,256,32,0.009305777649084726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,2048,0.033266666862699724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,65536,0.4038248856862386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,65536,1.0871707068549263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,768,0.016623111234770883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,16384,0.10336444775263469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,16384,0.27826399273342556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,1536,0.02618666655487484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,12288,0.07136178016662598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,12288,0.21170577738020155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,1024,0.01949155496226417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,10240,0.06012533108393351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,10240,0.1782595581478543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,512,0.013576000101036496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,8192,0.04832888974083794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,8192,0.14265067047542995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,256,0.010633777413103314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,7168,0.04310844341913859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,7168,0.12590489122602674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,256,128,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,6144,0.03872088922394647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,6144,0.10897599988513523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,65536,0.8030231263902453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,5120,0.03722755445374383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,5120,0.09458311398824056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,16384,0.20801244841681588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,4096,0.03035022152794732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,4096,0.07459288835525513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,12288,0.15928533342149523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,3584,0.023927999867333308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,3584,0.065792891714308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,10240,0.1334844430287679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,3072,0.021293333835071985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,3072,0.05653866794374254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,8192,0.10752355390124851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,2560,0.01837688850031959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,2560,0.04699466625849406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,7168,0.09488888581593831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,2048,0.015182221929232279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,6144,0.08239555358886719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,2048,0.034322665797339544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,1536,0.011853333148691388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,1536,0.025125333004527624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,5120,0.06891910897360908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,1024,0.008978666530715095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,1024,0.01865511139233907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,4096,0.05565333366394043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,768,0.007703999678293864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,768,0.018207111292415194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,3584,0.049329777558644615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,512,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,512,0.01273333364062839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,3072,0.04287733303176033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,256,0.003992889076471329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,256,0.009956444303194681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,2560,0.036658667855792575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,128,0.0035991109907627106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,128,0.008546666966544257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,64,0.00332266671790017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,64,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,128,32,0.003641777982314428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,128,32,0.008748444418112436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,65536,0.36601866616143125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,65536,1.1079208585951064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,16384,0.1057422227329678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,16384,0.28154132101270885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,12288,0.06918578015433417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,12288,0.21361511283450654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,10240,0.058182219664255776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,10240,0.17845600181155732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,8192,0.047285334931479565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,8192,0.14373600482940674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,7168,0.04191822144720289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,7168,0.12570488452911377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,6144,0.037307557132509016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,6144,0.10932799842622544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,5120,0.03143644332885742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,5120,0.09186222155888875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,4096,0.025972444150182936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,4096,0.07478400071461995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,3584,0.02309777836004893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,3584,0.06630400154325697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,3072,0.020463999774720933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,3072,0.0571848882569207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,2560,0.01763911048571269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,2560,0.047169778082105845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,2048,0.014520888527234396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,2048,0.03490666548411051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,1536,0.011537777880827585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,1536,0.025290666355027094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,1024,0.008623999853928884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,1024,0.019117333822780185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,768,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,768,0.018048000004556444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,512,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,512,0.012630222572220696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,256,0.0036968886852264404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,256,0.009760888914267222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,128,0.003321777822242843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,128,0.008432889150248634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,64,0.0031928889867332247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,64,0.008535111116038429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,64,32,0.0032604444358083936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,64,32,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,65536,0.3582879967159695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,65536,1.110838254292806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,16384,0.10477866729100545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,16384,0.28342223167419434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,12288,0.06830310821533203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,12288,0.21375021669599745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,10240,0.057387557294633657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,10240,0.1794346703423394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,8192,0.046691556771596275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,8192,0.14383467038472494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,7168,0.04143644372622172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,7168,0.1261342234081692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,6144,0.03668533431159125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,6144,0.10937955644395615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,5120,0.03094489044613308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,5120,0.09132799837324355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,4096,0.025389333566029865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,4096,0.07484266493055555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,3584,0.022633777724372015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,3584,0.06682311164008246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,3072,0.019928000039524503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,3072,0.05716088745329115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,2560,0.017129777206314933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,2560,0.04712444543838501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,2048,0.014053333136770459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,2048,0.03461066550678677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,1536,0.011223110887739392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,1536,0.02593244446648492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,1024,0.008466666771305932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,1024,0.019285332825448778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,768,0.007158222297827403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,768,0.01815555493036906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,512,0.0057795556883017225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,512,0.012642666697502136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,256,0.0034622223012977173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,256,0.009766221874290043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,128,0.003132444495956103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,128,0.008466666771305932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,64,0.002874666617976295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,64,0.008544888761308458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8192,32,32,0.00295733329322603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8192,32,32,0.008604444563388824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,16384,6.386054992675781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,16384,13.328792148166231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,16384,9.348263210720486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,12288,4.828691694471571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,12288,10.418504503038195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,12288,6.691105736626519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,10240,4.195753733317058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,10240,7.991104973687066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,10240,5.906917148166233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,8192,3.248990164862739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,8192,6.179798973931207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,8192,4.850941128200955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,7168,2.9384827084011502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,7168,5.426544189453125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,7168,3.652798122829861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,6144,2.535586675008138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,6144,4.668059455023871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,6144,2.782823986477322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,5120,2.1074657440185547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,5120,3.9032257927788625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,5120,2.2542008293999567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,4096,1.6724168989393446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,4096,3.11735471089681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,4096,1.7617244720458984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,3584,1.469950252109104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,3584,2.7023039923773875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,3584,1.5358880360921223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,3072,1.2744248708089192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,3072,2.345142152574327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,3072,1.3163430955674913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,2560,1.0442444483439128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,2560,2.070407019721137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,2560,1.0765946706136067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,2048,0.8018106884426541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,2048,1.6999048656887477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,2048,0.8673466576470269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,1536,0.6160453160603842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,1536,1.2532836066351998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,1536,0.6478764216105143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,1024,0.4496791097852919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,1024,0.8031368785434299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,1024,0.47893333435058594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,768,0.3639270994398329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,768,0.6057982444763184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,768,0.3894960085550944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,512,0.276692443423801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,512,0.4414462248484294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,512,0.34651022487216526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,256,0.24069423145718047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,256,0.276713768641154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,256,0.24775377909342447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,65536,128,0.2303911050160726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,128,0.22485955556233725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,128,0.23496622509426543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,64,0.20833333333333334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,64,0.2408320109049479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,65536,32,0.20345244142744276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,65536,32,0.2485884560479058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,65536,6.566718207465278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,65536,13.082776387532553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,65536,11.152995639377169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,16384,1.6952844195895727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,16384,3.1412648095024953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,16384,1.913368860880534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,12288,1.2228577931722004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,12288,2.4769778781467013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,12288,1.429483519660102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,10240,0.9715556038750542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,10240,2.0287208557128906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,10240,1.2912569046020508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,8192,0.8046995798746744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,8192,1.6748986774020727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,8192,0.9105964236789279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,7168,0.6941493352254232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,7168,1.4678497314453125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,7168,0.792135132683648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,6144,0.5848568810356988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,6144,1.2091644075181749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,6144,0.6675208939446343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,5120,0.4943777720133464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,5120,0.9794782002766927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,5120,0.5477395587497288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,4096,0.397582213083903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,4096,0.7559688886006674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,4096,0.44518131679958767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,3584,0.35216177834404844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,3584,0.6902355617947049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,3584,0.3942720095316569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,3072,0.32629866070217556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,3072,0.6080764664543999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,3072,0.33476887808905703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,2560,0.26002489195929634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,2560,0.4908409118652344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,2560,0.278111113442315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,2048,0.21435999870300293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,2048,0.38061777750651044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,2048,0.22323645485772026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,1536,0.1686542166603936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,1536,0.2896924548678928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,1536,0.1689795520570543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,1024,0.12321421835157607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,1024,0.1953751113679674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,1024,0.11882400512695312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,768,0.10016622145970662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,768,0.1537626716825697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,768,0.09910933176676433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,512,0.07655377520455255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,512,0.11314488781823052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,512,0.08230400085449219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,256,0.05892178085115221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,256,0.06979111168119642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,256,0.06859288613001506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,16384,128,0.05843822161356608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,128,0.060918225182427295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,128,0.06405688656700982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,64,0.056204444832272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,64,0.06610577636294894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,16384,32,0.055713779396480985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,16384,32,0.06744533114963107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,65536,5.019517262776692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,65536,10.588658650716146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,65536,7.3190502590603295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,16384,1.2247048483954537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,2048,0.030287110143237646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,16384,2.3583511776394315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,16384,1.383795526292589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,12288,1.794830322265625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,12288,1.1315146552191842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,12288,0.8799884584214953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,10240,1.4858328501383464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,10240,0.9200106726752387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,10240,0.726818667517768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,8192,1.1854995091756184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,8192,0.6761982176038953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,8192,0.595254209306505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,7168,1.0162827173868816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,7168,0.5967137548658583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,7168,0.5159164534674751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,6144,0.8954426447550455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,6144,0.5225937631395128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,5120,0.4140053325229221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,5120,0.7558195326063367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,5120,0.42916356192694766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,6144,0.4459182421366374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,4096,0.555298646291097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,4096,0.3566293451521132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,4096,0.30250665876600474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,3584,0.47159020105997723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,3584,0.3007928795284695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,3584,0.269503116607666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,3072,0.4199662208557129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,3072,0.25805332925584584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,3072,0.23159289360046387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,2560,0.339300447040134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,2560,0.21701333257887098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,2560,0.199325336350335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,2048,0.2758222156100803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,2048,0.17188799381256104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,2048,0.16370044814215765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,1536,0.20970577663845488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,1536,0.13113955656687418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,1536,0.1297031111187405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,1024,0.14770578013526067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,1024,0.09315377473831177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,1024,0.09657600190904404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,768,0.11594488885667588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,768,0.07764799727333917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,768,0.07799022065268622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,512,0.08580355511771308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,512,0.06568088796403673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,512,0.061499555905659996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,256,0.053912887970606484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,256,0.05411377880308363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,256,0.04733777708477444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,128,0.04751377635531955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,128,0.050612443023257785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,64,0.04372444417741564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,64,0.051960001389185585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,12288,32,0.043545779254701406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,12288,32,0.05273155702484978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,12288,128,0.046457777420679726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,65536,4.238079071044922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,65536,7.698789808485244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,65536,5.830087873670791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,16384,0.9833466211954752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,16384,2.247392018636068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,16384,1.1789866553412545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,12288,0.9218311309814453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,12288,0.9284391403198242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,12288,1.506058692932129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,10240,1.1919519636366103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,10240,0.7609350946214465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,10240,0.6213368839687771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,8192,1.0299253463745117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,8192,0.5857564608256022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,8192,0.5025528801812066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,7168,0.8305564456515842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,7168,0.5064364539252387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,7168,0.434700436062283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,6144,0.7202346589830187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,6144,0.44004355536566836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,6144,0.37700711356268984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,5120,0.573760880364312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,5120,0.3633982340494792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,5120,0.314717345767551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,4096,0.47890133327907985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,4096,0.296378665500217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,4096,0.25711554951137966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,3584,0.4137004481421576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,3584,0.25782399707370335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,3584,0.22692354520161948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,3072,0.3437679873572455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,3072,0.21807644102308485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,3072,0.19727910889519584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,2560,0.2916853427886963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,2560,0.18190133571624756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,2560,0.1680177715089586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,2048,0.23507645395067003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,2048,0.1397093269560072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,2048,0.14611377980973986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,1536,0.11026933458116318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,1536,0.18722666634453666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,1536,0.11131200525495742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,1024,0.1264924473232693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,1024,0.07941422197553846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,1024,0.08186844322416517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,768,0.06755999724070232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,768,0.06670222017500135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,768,0.09970133172141181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,512,0.07412888606389363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,512,0.05624533361858792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,512,0.05221866567929586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,256,0.0499857763449351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,256,0.04699288805325826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,256,0.04028622309366862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,128,0.04069066709942288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,128,0.04331644376118978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,64,0.036610666248533465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,64,0.04455644554562039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,10240,32,0.0368115570810106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,10240,32,0.0451182226339976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,65536,3.445335176255968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,10240,128,0.039509332842297025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,65536,6.237125396728516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,65536,4.652468363444011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,16384,1.4951360490587022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,16384,1.0057769351535373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,16384,0.803610642751058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,12288,1.1410915586683485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,12288,0.8068559964497884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,12288,0.6031662093268501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,10240,0.9365982479519315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,10240,0.6122888988918728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,10240,0.500399112701416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,8192,0.7695751190185547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,8192,0.4746053483751085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,8192,0.40537065929836696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,7168,0.6348124610053169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,7168,0.4219200081295437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,7168,0.35540178087022567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,6144,0.5379021962483724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,6144,0.36073509852091473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,6144,0.30736621220906574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,5120,0.4522293408711751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,5120,0.2975848780737983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,5120,0.25724267959594727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,4096,0.36047999064127606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,4096,0.2100435495376587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,4096,0.23674223158094618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,3584,0.31517601013183594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,3584,0.21095732847849527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,3584,0.186337775654263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,3072,0.2741413381364611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,3072,0.17948355939653185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,2560,0.13881066110399035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,2560,0.2308328946431478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,2560,0.15057778358459473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,3072,0.16062488820817736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,2048,0.18485421604580352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,2048,0.12071466445922852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,2048,0.11461067199707031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,1536,0.14261421892378065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,1536,0.09273599916034275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,1536,0.09103822045856053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,1024,0.09977600309583877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,1024,0.06397689051098294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,1024,0.06635733445485432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,768,0.07975822024875216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,768,0.054720001088248364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,768,0.054719110329945884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,512,0.05855466922124227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,512,0.046366221374935575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,512,0.041548444165123835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,256,0.03801244497299194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,256,0.038605332374572754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,128,0.03202399942609999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,128,0.03344444433848063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,128,0.0359208881855011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,64,0.029598222838507757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,64,0.036743998527526855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,8192,32,0.030376000536812678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,8192,32,0.037137776613235474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,8192,256,0.03279555506176419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,65536,5.353799184163411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,65536,3.0143422020806203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,65536,4.057342105441623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,16384,1.3065129386054146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,16384,0.8932257758246528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,16384,0.7168364524841309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,12288,0.9759972890218099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,12288,0.651476436191135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,12288,0.5342986848619249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,10240,0.816432899898953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,10240,0.5548293325636122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,10240,0.4479377534654405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,8192,0.6508560180664062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,8192,0.42264800601535374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,8192,0.35935645633273655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,7168,0.566447999742296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,7168,0.37162134382459855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,6144,0.31715554661220974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,6144,0.51346312628852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,6144,0.3346889019012451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,7168,0.3181324534946018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,5120,0.3967128859625922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,5120,0.26641776826646596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,5120,0.22917599148220488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,4096,0.31417422824435765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,4096,0.20994399653540718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,4096,0.18740444713168672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,3584,0.27828976843092174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,3584,0.18590577443440756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,3584,0.1654320028093126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,3072,0.23885867330763075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,3072,0.15928088294135198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,3072,0.14428267214033338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,2560,0.20155378182729086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,2560,0.133769777086046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,2560,0.12334844801161025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,2048,0.16254577371809217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,2048,0.10771022240320842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,2048,0.10250044531292385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,1536,0.1254862282011244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,1536,0.08278755346934001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,1536,0.08156355884340075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,1024,0.08839999967151219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,1024,0.05794666873084175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,1024,0.060039109653896756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,768,0.07018755541907416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,768,0.04973333411746555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,768,0.04965777860747444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,512,0.05193422238032023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,512,0.042846222718556724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,512,0.03828977876239353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,256,0.03388088941574097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,256,0.03506933318244086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,256,0.030228445927302044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,128,0.030065778228971694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,128,0.03241422110133701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,64,0.026838221483760413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,64,0.03362844387690226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,7168,32,0.02780533168050978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,7168,32,0.03389777739842733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,7168,128,0.02877955635388692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,65536,4.596309238009982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,65536,3.3217671712239585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,65536,2.6479049258761935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,16384,1.1308026843600802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,16384,0.7417244381374783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,16384,0.6220453580220541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,12288,0.8472168710496691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,12288,0.5593991279602051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,10240,0.43928888108995223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,10240,0.7330337630377876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,10240,0.4755688773261176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,12288,0.4676355785793728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,8192,0.5706746843126085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,8192,0.3660764429304335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,8192,0.3143262333340115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,7168,0.46572086546156144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,7168,0.3268871042463514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,7168,0.2751884460449219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,6144,0.4017048941718207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,6144,0.28299466768900555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,6144,0.2386239899529351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,5120,0.33568443192376035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,5120,0.23285155826144746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,5120,0.20010577307807076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,4096,0.2691360049777561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,4096,0.18581688404083252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,4096,0.16362578339046902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,3584,0.23845688501993814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,3584,0.1635271045896742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,3584,0.14469067255655924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,3072,0.2047182189093696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,3072,0.13973244031270346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,3072,0.12619466251797146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,2560,0.1732391119003296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,2560,0.11726578076680501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,2560,0.10877244340048896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,2048,0.1394408941268921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,2048,0.09460800223880345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,2048,0.09042133225335015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,1536,0.10829777850045098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,1536,0.07238044341405232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,1536,0.07142400079303317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,1024,0.07618221971723768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,1024,0.050923556089401245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,1024,0.053017778529061206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,768,0.061066667238871254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,768,0.04419822163052029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,768,0.04390933447413974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,512,0.04536711176236471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,512,0.03814399904674954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,512,0.034507556094063654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,256,0.03018222252527873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,256,0.03137333194414774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,128,0.025796444879637823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,128,0.026134222745895386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,128,0.029101332028706867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,64,0.023863111933072407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,64,0.02986933456526862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,6144,32,0.024494222468800012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,6144,32,0.029821332958009508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,6144,256,0.027686221732033625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,65536,3.8571963840060763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,65536,2.3497272067599826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,65536,2.843639161851671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,16384,0.946290651957194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,16384,0.6417528788248698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,16384,0.5278462304009331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,12288,0.6990622414482964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,12288,0.4845680130852594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,12288,0.39823200967576766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,10240,0.5845102204216851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,10240,0.4099591043260362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,10240,0.3341137833065457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,8192,0.44799110624525285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,8192,0.3169146643744575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,8192,0.2680933210584852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,7168,0.39748088518778485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,7168,0.2819306585523817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,7168,0.23748355441623262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,6144,0.3415333429972331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,6144,0.24325156211853027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,5120,0.1863422261344062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,5120,0.2918773227267795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,5120,0.20005510913001168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,6144,0.2037688891092936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,4096,0.2278266747792562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,4096,0.15935199790530735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,3584,0.1337946653366089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,3584,0.20183111561669242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,3584,0.1406595574484931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,4096,0.14047200149959987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,3072,0.17462310526106092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,3072,0.11955466535356309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,3072,0.10858400000466241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,2560,0.14671999878353542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,2560,0.10079199737972683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,2560,0.09347110986709595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,2048,0.12013333373599583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,2048,0.0804248915778266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,2048,0.07767733600404528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,1536,0.09243110815684001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,1536,0.06210933128992716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,1536,0.06154221958584256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,1024,0.0463084446059333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,1024,0.0654311113887363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,1024,0.04416444566514757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,768,0.05261066555976868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,768,0.03910577628347609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,768,0.03810133205519782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,512,0.04003644320699904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,512,0.03307377629809909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,256,0.023823999696307715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,256,0.028064890040291682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,256,0.027624888552559748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,512,0.030232889784706965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,128,0.022310222188631695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,128,0.025184889634450276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,64,0.019718221492237516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,64,0.02588622272014618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,5120,32,0.02029333346419864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,5120,32,0.025993777645958796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,5120,128,0.02235466738541921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,65536,1.8401448991563585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,65536,3.2500587039523654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,65536,2.0725252363416886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,16384,0.744509326087104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,16384,0.49668264389038086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,16384,0.43515555063883465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,12288,0.5681849055820042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,12288,0.3974017832014296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,12288,0.32790933714972603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,10240,0.44195466571384007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,10240,0.32649866739908856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,10240,0.27340533998277455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,8192,0.3524888886345758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,8192,0.25049156612820095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,7168,0.21107821994357637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,7168,0.32643289036220974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,7168,0.22735823525322807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,8192,0.22227377361721465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,6144,0.26839733123779297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,6144,0.19314400355021158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,6144,0.1681742270787557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,5120,0.2232266796959771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,5120,0.15554932753245035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,5120,0.1427288850148519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,4096,0.1818924480014377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,4096,0.12141244941287571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,4096,0.11656978395250107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,3584,0.159116440349155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,3584,0.10087111261155869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,3584,0.1040782199965583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,3072,0.1390497816933526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,3072,0.08465866910086738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,2560,0.07858310805426703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,2560,0.11777067184448242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,2560,0.07394399907853869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,3072,0.09052711062961155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,2048,0.09501689010196263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,2048,0.06173333194520739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,1536,0.0520648890071445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,1536,0.07366933425267537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,1536,0.048680888281928174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,2048,0.06553510824839275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,1024,0.05272000034650167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,1024,0.036666668123669095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,1024,0.03938400083118015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,768,0.04244088795449999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,768,0.03257511059443156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,768,0.03248622351222568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,512,0.03219377663400438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,512,0.026495110657480028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,512,0.025763556361198425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,256,0.022646221849653456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,256,0.02202133337656657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,256,0.020415999823146395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,128,0.017704889178276062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,128,0.01958311100800832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,64,0.014910222755538093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,64,0.020035554965337116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,4096,32,0.015669332610236276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,4096,32,0.020473778247833252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,4096,128,0.01903377804491255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,65536,1.8262924618191188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,65536,2.793430116441515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,65536,1.9662329355875652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,16384,0.7118817965189616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,16384,0.4511120054456923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,16384,0.43326489130655926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,12288,0.4965039889017741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,12288,0.35066133075290257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,12288,0.32900089687771267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,10240,0.4198577668931749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,10240,0.2953502337137858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,10240,0.2746720049116346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,8192,0.3277795579698351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,8192,0.22882045639885795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,8192,0.22269778781467012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,7168,0.29000621371799046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,7168,0.20216800106896293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,7168,0.1959448920355903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,6144,0.2481386661529541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,6144,0.1744124492009481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,6144,0.17006666130489775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,5120,0.14405243926578098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,5120,0.2096551126903958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,5120,0.14252889156341553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,4096,0.16829066806369355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,4096,0.11240443918440078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,4096,0.11793333954281277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,3584,0.14962932798597547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,3584,0.0951733324262831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,3584,0.10574755403730605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,3072,0.12851999865637884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,3072,0.08179644743601482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,3072,0.09155378076765273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,2560,0.10969866646660699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,2560,0.07112089130613539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,2560,0.0793351133664449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,2048,0.08873511022991604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,2048,0.05768711037105984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,2048,0.06586488750245836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,1536,0.06933955351511638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,1536,0.04483911063936022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,1536,0.052998221582836576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,1024,0.0495671100086636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,1024,0.033969776497946844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,1024,0.04161422120200263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,768,0.04039999842643738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,768,0.03005689051416185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,768,0.03387199838956197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,512,0.030188444587919448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,512,0.024459555745124817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,512,0.027652444110976324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,256,0.02072355482313368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,256,0.02020711037847731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,256,0.021319111188252766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,128,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,128,0.018070222602950204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,64,0.014202665951516895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,64,0.018183110488785636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3584,32,0.015632889337009855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3584,32,0.017927999297777813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,65536,1.4322808583577473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3584,128,0.01893422173129188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,65536,2.2770631578233504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,65536,1.6314915551079645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,16384,0.5662942462497288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,16384,0.40613423453436953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,16384,0.3451697826385498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,12288,0.3938417699601915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,12288,0.3109102249145508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,12288,0.26133423381381565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,10240,0.3313475449879964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,10240,0.26075998942057294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,10240,0.21930844253963896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,8192,0.2644897831810845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,8192,0.21162133746676973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,8192,0.17805333932240805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,7168,0.23417954974704316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,7168,0.17976178063286674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,7168,0.15658044815063477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,6144,0.2004826731152005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,6144,0.1551626655790541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,6144,0.13569066259596083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,5120,0.16940800348917642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,5120,0.12629066573248968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,5120,0.11492355664571126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,4096,0.13620355394151476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,4096,0.09866222408082749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,4096,0.09408799807230632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,3584,0.12132622136010064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,3584,0.08421778016620213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,3584,0.08394489023420547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,3072,0.10433155298233032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,3072,0.0731875565316942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,3072,0.07347822189331055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,2560,0.08910933468076919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,2560,0.06342399782604642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,2560,0.06377955277760823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,2048,0.07235466771655612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,2048,0.05149955550829569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,2048,0.053301334381103516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,1536,0.05645155244403415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,1536,0.04019200139575534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,1536,0.04221333397759331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,1024,0.040644443697399564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,1024,0.030442665020624798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,1024,0.03219022353490194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,768,0.03315022256639268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,768,0.02714666724205017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,768,0.027015111512607996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,512,0.02551288902759552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,512,0.021781333618693884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,512,0.02223199937078688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,256,0.018408889571825664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,256,0.018059555027219985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,128,0.015951111912727356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,128,0.014351111319330005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,128,0.016057777735922072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,64,0.01221333361334271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,64,0.0159599996275372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,3072,32,0.012632889052232107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,3072,32,0.01609511176745097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,3072,256,0.017482666505707633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,65536,1.815142313639323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,65536,1.4399439493815105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,65536,1.4554275936550563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,16384,0.4446266492207845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,16384,0.3564559883541531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,16384,0.34538666407267254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,12288,0.32496534453497994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,12288,0.276455110973782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,12288,0.26324354277716744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,10240,0.2730071014828152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,10240,0.23142843776279023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,10240,0.22049066755506727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,8192,0.2181884447733561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,8192,0.17967555258009169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,8192,0.17926222748226592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,7168,0.1930711136923896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,7168,0.1589804490407308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,6144,0.14102399349212646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,6144,0.16675021913316515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,6144,0.13745510578155518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,7168,0.15763289398617214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,5120,0.1394773324330648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,5120,0.11255466938018799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,5120,0.1159564389122857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,4096,0.11289600531260173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,4096,0.08757688601811726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,4096,0.09458933273951213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,3584,0.09884177976184422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,3584,0.07656444443596734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,3584,0.08420888582865398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,3072,0.08629333310657078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,3072,0.06562755505243938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,3072,0.07429866658316718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,2560,0.07256711191601224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,2560,0.05719466341866387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,2560,0.06413155794143677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,2048,0.05968088573879666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,2048,0.04605422086185879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,2048,0.05386044581731161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,1536,0.04632622334692213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,1536,0.03614577651023865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,1536,0.0432168874475691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,1024,0.032904889848497175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,1024,0.027434666951497395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,1024,0.03415822320514255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,768,0.026353778110610113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,768,0.024400000770886738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,768,0.028265777561399672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,512,0.020044444335831534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,512,0.019544000426928203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,512,0.023231999741660223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,256,0.014196443888876172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,256,0.01608533329433865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,256,0.017447110679414537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,128,0.010631110933091907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,128,0.014076444837782117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,64,0.009087111386987898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,64,0.01420177850458357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2560,32,0.009291555318567488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2560,32,0.014355555176734924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2560,128,0.015619556109110514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,65536,1.4390701717800563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,65536,1.0337804158528645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,65536,1.2164426379733615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,16384,0.3579386605156793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,16384,0.3030782275729709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,12288,0.20321244663662383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,12288,0.2625848982069227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,12288,0.2390026781294081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,16384,0.25645332866244847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,10240,0.22033954991234672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,10240,0.19915378093719482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,8192,0.1372319989734226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,8192,0.17764622635311553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,8192,0.15607199403974745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,10240,0.16327555974324545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,7168,0.15455822149912515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,7168,0.13782310485839844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,7168,0.1171262264251709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,6144,0.13418755266401503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,6144,0.11954666508568658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,6144,0.10124178065194024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,5120,0.11165777842203777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,5120,0.09865422381295098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,5120,0.08637777964274089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,4096,0.09112711085213555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,4096,0.07719377676645915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,4096,0.07134666707780626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,3584,0.08039466540018718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,3584,0.06713066498438518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,3072,0.05658400058746338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,3072,0.06920266813702054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,3072,0.05754577451282077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,3584,0.06419999731911553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,2560,0.05877955754597982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,2560,0.04942577746179369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,2048,0.04105866617626614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,2048,0.048335111803478666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,2560,0.049448887507120766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,2048,0.040403554836908974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,1536,0.03748177819781833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,1536,0.031368888086742826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,1536,0.032885332902272545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,1024,0.026649778087933857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,1024,0.02366311185889774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,1024,0.025011556016074285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,768,0.021372444099850122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,768,0.02164977788925171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,768,0.02158311174975501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,512,0.01668444441424476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,512,0.01735822194152408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,512,0.017813333206706576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,256,0.011889778077602386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,256,0.014031110538376702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,256,0.014180445008807711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,128,0.009113777842786577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,128,0.012015110916561551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,64,0.007868444754017724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,64,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,2048,32,0.008032889001899296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,2048,32,0.012485333614879184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,2048,128,0.012888000243239932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,65536,0.8349137836032443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,65536,1.10499996609158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,16384,0.23700533972846138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,16384,0.2875102096133762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,65536,1.0669022666083443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,16384,0.26275732782151967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,12288,0.2114035553402371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,12288,0.2011528942320082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,12288,0.15942222542232937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,10240,0.16753866937425402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,10240,0.13406399885813394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,10240,0.17083999845716688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,8192,0.13519377178615993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,8192,0.1325706640879313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,8192,0.1083822250366211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,7168,0.1193431086010403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,7168,0.11766578091515435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,7168,0.09613777531517877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,6144,0.10232710838317871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,6144,0.10218133529027303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,6144,0.0839360025193956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,5120,0.08659555514653523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,5120,0.08515022198359172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,5120,0.0711359977722168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,4096,0.0694906645350986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,4096,0.06541866726345487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,4096,0.05907555421193441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,3584,0.06201777855555216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,3584,0.05764000283347236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,3584,0.05316177672810025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,3072,0.05335200164053175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,3072,0.049478222926457725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,3072,0.04742133286264208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,2560,0.04533333248562283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,2560,0.0425600012143453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,2560,0.04085866610209147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,2048,0.03693244523472256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,2048,0.03420177764362759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,2048,0.034020443757375084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,1536,0.028805332051383123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,1536,0.026848889059490625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,1536,0.02756355537308587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,1024,0.02056533263789283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,1024,0.02089599933889177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,1024,0.02164177762137519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,768,0.016943110360039603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,768,0.019083556201722886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,512,0.015786666009161208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,512,0.013407111167907715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,512,0.015125332607163323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,768,0.018768888380792405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,256,0.009287111461162567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,256,0.011688888900809817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,256,0.012513777448071374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,128,0.007228444847795699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,128,0.01037333326207267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,64,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,64,0.01075466639465756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1536,32,0.006569777925809224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1536,32,0.010813333094120026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1536,128,0.011074666347768573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,65536,0.7184826532999674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,65536,0.8544400003221301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,65536,0.6673697895473905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,16384,0.17572444015079072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,16384,0.2159271107779609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,16384,0.16980177826351592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,12288,0.1317395501666599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,12288,0.16575911309983996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,12288,0.12880888250139025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,10240,0.1112942165798611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,10240,0.13811288939581975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,10240,0.10856266816457112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,8192,0.0890702207883199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,8192,0.10953866773181492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,7168,0.07860977782143487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,7168,0.0783546699417962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,7168,0.09702133470111424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,8192,0.08867288960350884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,6144,0.06836089160707262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,6144,0.08497689167658488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,6144,0.06893599695629544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,5120,0.057103998131222196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,5120,0.06990133391486274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,5120,0.05915021896362305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,4096,0.046635554896460645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,4096,0.054986665646235146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,4096,0.04951111144489712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,3584,0.04109333289994134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,3584,0.04792799883418613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,3584,0.04452355702718099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,3072,0.035981333918041654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,3072,0.04099022348721822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,3072,0.03899733225504557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,2560,0.030388444662094116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,2560,0.035036444664001465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,2560,0.033497777250077986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,2048,0.025224000215530396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,2048,0.02873777680926853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,2048,0.02814488940768772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,1536,0.019871999820073444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,1536,0.023043556345833674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,1536,0.023230221536424425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,1024,0.014992000328169929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,1024,0.018408889571825664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,1024,0.018577777677112155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,768,0.012583110895421771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,768,0.016664889123704698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,768,0.016190222567982145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,512,0.01018133345577452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,512,0.01273155543539259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,512,0.0140364451540841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,256,0.007300444775157505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,256,0.009673777553770278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,256,0.01144444445768992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,128,0.006051555689838197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,128,0.008775999976529015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,64,0.005370666583379109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,64,0.00907022257645925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,1024,32,0.005535999933878581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,1024,32,0.009118222528033787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,1024,128,0.010090666512648264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,65536,0.5445413589477539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,65536,0.8295146624247233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,65536,0.578038215637207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,16384,0.13553688261244032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,16384,0.2085955540339152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,16384,0.14665067195892334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,12288,0.10162311130099827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,12288,0.1582497755686442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,12288,0.11139555772145589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,10240,0.08598133590486313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,10240,0.13240267170800105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,8192,0.07670044236712985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,8192,0.06945511367585924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,8192,0.10463377502229478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,10240,0.0937911139594184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,7168,0.06063555346594917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,7168,0.09318311346901788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,7168,0.0681848857137892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,6144,0.05285066697332594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,6144,0.08179644743601482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,6144,0.05968177980846829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,5120,0.044212443961037524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,5120,0.06755022207895915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,5120,0.0511813329325782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,4096,0.03623022304640876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,4096,0.052968889474868774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,4096,0.04297688934538099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,3584,0.031926221317715116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,3584,0.045533332559797496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,3072,0.03388799892531501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,3072,0.027874665127860174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,3072,0.03888088795873854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,3584,0.038471112648646034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,2560,0.023904000719388325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,2560,0.03307466705640157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,2048,0.024450666374630396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,2048,0.019786667492654588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,2048,0.027111111415757075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,2560,0.029191109869215224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,1536,0.01595288846227858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,1536,0.022395556171735127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,1536,0.020557334025700886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,1024,0.012073777616024017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,1024,0.017474666237831116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,1024,0.01665866706106398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,768,0.010302222437328762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,768,0.016322667400042217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,768,0.014452444182501899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,512,0.008271111382378472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,512,0.012677333421177335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,512,0.012772444221708508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,256,0.006051555689838197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,256,0.009510222408506606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,256,0.010473777850468954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,128,0.005110222018427319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,128,0.00870400004916721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,64,0.004575110971927643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,64,0.008847110801272923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,768,32,0.004805333498451444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,768,32,0.008893333375453949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,768,128,0.009062222308582729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,65536,0.3690933386484782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,65536,0.6510720252990723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,65536,0.5195306671990289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,16384,0.09334755606121486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,16384,0.16512800587548151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,16384,0.13341066572401258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,12288,0.06973333491219415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,12288,0.12683822049034968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,12288,0.10259377956390381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,10240,0.05891999933454725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,10240,0.1064337756898668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,10240,0.0868871079550849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,8192,0.047116445170508496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,8192,0.0853946672545539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,8192,0.06860799921883477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,7168,0.04198044538497925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,7168,0.07529244158003065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,7168,0.06152444415622287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,6144,0.036185777849621244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,6144,0.06575822167926364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,6144,0.054697778489854604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,5120,0.03089155422316657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,5120,0.05476266807980008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,5120,0.04664355516433716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,4096,0.02516533268822564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,4096,0.042239109675089516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,4096,0.03834755553139581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,3584,0.022621333599090576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,3584,0.036438223388459944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,3584,0.03467911150720384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,3072,0.019581332802772522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,3072,0.03086044390996297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,3072,0.030088888274298772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,2560,0.01687733332316081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,2560,0.026557332939571802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,2560,0.025940444734361436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,2048,0.014137778017255994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,2048,0.02216533323129018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,2048,0.02183377742767334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,1536,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,1536,0.018459555175569322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,1536,0.018384888768196106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,1024,0.0092275556590822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,1024,0.014764444695578681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,1024,0.014599111345079211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,768,0.008050666915045844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,768,0.013917333549923368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,768,0.012930666406949362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,512,0.006477333191368315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,512,0.010642666783597736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,512,0.01126933346192042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,256,0.004912000149488449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,256,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,256,0.009023110899660323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,128,0.004344888859324985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,128,0.00741244438621733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,64,0.003986666599909465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,64,0.00756444450881746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,512,32,0.004173333446184794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,512,32,0.007674666742483775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,512,128,0.0074240002367231585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,65536,0.21808000405629477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,65536,0.6042444441053603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,65536,0.4494471020168728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,16384,0.06554933389027913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,16384,0.15579022301567927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,16384,0.11511822541554768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,12288,0.049572444624371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,12288,0.119504001405504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,12288,0.08930222193400066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,10240,0.04492888847986857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,10240,0.10037510924869114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,10240,0.07674044370651245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,8192,0.03549422158135308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,8192,0.08095555835300021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,7168,0.054289778073628746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,7168,0.02933244572745429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,7168,0.07185244560241699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,8192,0.06098666456010607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,6144,0.02864177690611945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,6144,0.062478224436442055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,6144,0.048138668139775596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,5120,0.02423911127779219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,5120,0.05194044444296095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,5120,0.0413351125187344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,4096,0.019696000549528334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,4096,0.03923822111553616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,4096,0.034323556555642024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,3584,0.01628800067636702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,3584,0.03422044383154975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,3584,0.030604445272021826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,3072,0.01519377695189582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,3072,0.029085334804322984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,3072,0.02671555512481266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,2560,0.012305777933862476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,2560,0.02514755560292138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,2560,0.02327022287580702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,2048,0.01074044406414032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,2048,0.021538666552967493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,1536,0.01643733349111345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,1536,0.008583111067612966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,1536,0.017808889349301655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,2048,0.019658666517999437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,1024,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,1024,0.014203555054134794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,1024,0.01313244468635983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,768,0.006124444305896759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,768,0.013688888814714221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,768,0.011218667030334473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,512,0.0047217777205838096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,512,0.01035733355416192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,512,0.009472000102202097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,256,0.003947555604908201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,256,0.007851555943489075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,256,0.006716444260544247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,128,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,128,0.007166222565703922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,64,0.003310222178697586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,64,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,256,32,0.0034862222770849862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,256,32,0.007366221812036302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,256,128,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,65536,0.19480621814727783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,65536,0.5939777692159017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,65536,0.45016177495320636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,16384,0.06172088781992594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,16384,0.1533706718020969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,16384,0.1124604410595364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,12288,0.04390133420626322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,12288,0.11798311604393853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,10240,0.0723137789302402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,10240,0.033816887272728816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,10240,0.09894311428070068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,12288,0.0857057770093282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,8192,0.027095999982621934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,8192,0.0799386633767022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,8192,0.058766219351026744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,7168,0.024506666594081458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,7168,0.0710924466451009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,7168,0.051911112334993154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,6144,0.021998221675554912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,6144,0.06145866711934408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,6144,0.04513955447408888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,5120,0.01902844508488973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,5120,0.05095022254519992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,5120,0.03841777642567953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,4096,0.01531733406914605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,4096,0.03814222084151374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,4096,0.031883554326163396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,3584,0.01387733303838306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,3584,0.03273155623012119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,3584,0.028374221589830186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,3072,0.012439111040698158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,3072,0.027879112296634253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,3072,0.02498666610982683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,2560,0.010861333045694562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,2560,0.024634665913052026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,2560,0.021469333105617102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,2048,0.009522666533788046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,2048,0.02144977781507704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,2048,0.018166222506099276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,1536,0.007903999752468532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,1536,0.01740622189309862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,1536,0.023872888750500146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,1024,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,1024,0.014000889327791003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,1536,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,768,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,768,0.013471110827393003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,1024,0.017658667431937326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,512,0.004057777838574516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,512,0.0098213329911232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,1024,0.011923555698659686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,256,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,256,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,768,0.014727999766667684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,128,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,128,0.007045333584149678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,64,0.003027555636233754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,64,0.007129777636792924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,128,32,0.0032382222513357797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,128,32,0.007108444141017065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,65536,0.18422310882144502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,65536,0.5964124467637804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,16384,0.06154133213890923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,16384,0.15306133694118923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,12288,0.04105688797103034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,12288,0.11710666285620795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,10240,0.034966223769717746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,10240,0.09851021899117364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,8192,0.029250668154822454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,8192,0.07955911424424914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,7168,0.0265964451763365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,7168,0.07084800137413873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,6144,0.02100177771515316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,6144,0.061490668190850147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,5120,0.018000000052981906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,5120,0.05052533414628771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,4096,0.014927110738224454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,4096,0.0381351113319397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,3584,0.013221333424250284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,3584,0.03286044465170966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,3072,0.011953777737087674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,3072,0.02797244323624505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,2560,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,2560,0.02446222139729394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,2048,0.009172444542249044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,2048,0.021412443783548143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,1536,0.007694222033023834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,1536,0.0173680004146364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,1024,0.006259555617968242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,1024,0.014077333940400017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,768,0.005032000028424793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,768,0.013549333645237817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,512,0.0037413334680928122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,512,0.009554666777451834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,256,0.0032915555768542816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,256,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,128,0.0030577776746617425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,128,0.00701333334048589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,64,0.0028977776981062363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,64,0.007080888582600488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,64,32,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,64,32,0.007080000307824876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,65536,0.18022399478488496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,65536,0.5988328721788194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,16384,0.05730399820539686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,16384,0.15271022584703234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,12288,0.03957510987917582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,12288,0.11686133013831244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,10240,0.033583111233181424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,10240,0.09825333621766831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,8192,0.02818044357829624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,8192,0.07953155703014798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,7168,0.02536444365978241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,7168,0.07069244649675158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,6144,0.022686221533351477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,6144,0.061457779672410756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,5120,0.01756444407833947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,768,0.010142222046852112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,5120,0.05075110991795858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,4096,0.014494222071435718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,4096,0.03799644443723891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,3584,0.01308444473478529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,3072,0.011759999725553723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,3584,0.0332275562816196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,2560,0.010270222193664974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,3072,0.02797244323624505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,2048,0.008840888738632202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,2560,0.0243493335114585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,1536,0.007517333659860823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,2048,0.021420444051424663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,1536,0.01721955504682329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,1024,0.006019555446174409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,768,0.00452977791428566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,1024,0.013993778162532382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,768,0.013478221992651621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,512,0.003448888866437806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,256,0.00309955556359556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,512,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,256,0.007676444119877285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,128,0.0028613333900769553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,128,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,64,0.002784888984428512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4096,32,32,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,64,0.0069857777820693115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4096,32,32,0.0070977773931291364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,16384,6.280440860324436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,16384,3.4942639668782554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,16384,3.339252471923828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,12288,4.964916653103299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,12288,3.1744302113850913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,12288,2.5478568606906467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,10240,4.0642009311252165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,10240,2.1682205200195312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,10240,2.1128436194525824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,8192,3.086666742960612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,8192,1.603851530287001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,8192,1.6923404269748266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,7168,2.7382835812038846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,7168,1.4957946141560872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,7168,1.4343751271565754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,6144,2.3055803510877824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,6144,1.2717875374688041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,5120,1.0066711637708876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,6144,1.2374497519599068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,5120,1.9402781592475042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,5120,0.9941280153062609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,4096,1.694175084431966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,4096,0.7940204408433703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,4096,0.7567875650193957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,3584,1.4664888381958008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,3584,0.687063111199273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,3584,0.6976862483554416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,3072,1.2015635172526042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,3072,0.5826586617363824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,3072,0.5787377887301975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,2560,1.0414799584282768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,2560,0.49265421761406797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,2560,0.49461687935723203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,2048,0.7673955493503146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,2048,0.40222933557298446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,2048,0.40877511766221786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,1536,0.5698106553819444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,1536,0.3122062153286404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,1536,0.3187928994496663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,1024,0.3914631207784017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,1024,0.22522576649983725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,1024,0.2331519921620687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,768,0.3068879975212945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,768,0.19196889135572645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,512,0.14436533715989855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,512,0.2270106739468045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,512,0.17379289203219941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,256,0.11424711015489365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,256,0.14849778016408285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,256,0.1387031078338623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,768,0.19209067026774088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,128,0.11689510610368516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,128,0.13128444883558485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,64,0.11616267098320855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,64,0.13262044058905706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,65536,32,0.11318222681681316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,65536,32,0.13424621687995061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,65536,128,0.11930133236779107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,65536,6.1033689710829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,65536,3.346377690633138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,65536,3.8897726270887585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,16384,1.9384746551513672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,16384,0.8748248947991265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,16384,0.7780639860365125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,12288,1.1899342007107205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,12288,0.6157795588175455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,12288,0.595045354631212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,10240,0.9663972854614258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,10240,0.524712880452474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,10240,0.4833822250366211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,8192,0.7729120254516602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,8192,0.4030471112993028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,7168,0.3765244483947754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,7168,0.6725342008802625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,7168,0.3660702175564236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,8192,0.38959111107720273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,6144,0.5361990928649902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,6144,0.32176534334818524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,5120,0.2705911000569661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,5120,0.4660862286885579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,5120,0.26197510295444065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,6144,0.2936577796936035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,4096,0.36648445659213597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,4096,0.2044311099582248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,4096,0.20035910606384277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,3584,0.3183777862124973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,3584,0.18899910979800752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,3584,0.1766453319125705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,3072,0.2732666598425971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,3072,0.15475911564297146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,3072,0.15354043907589382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,2560,0.2305306593577067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,2560,0.13452000088161892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,2560,0.13098843892415366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,2048,0.18551199966006807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,2048,0.10725777679019505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,1536,0.08816088570488824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,1536,0.14345955848693848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,1536,0.08389066987567478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,2048,0.1090444458855523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,1024,0.10132533311843872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,1024,0.06282044119305082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,768,0.054211556911468506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,768,0.08022044764624701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,768,0.054304000404145986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,1024,0.06450844473308988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,512,0.059304886394076876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,512,0.04855644371774462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,512,0.04105866617626614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,256,0.0384844442208608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,256,0.04190488987498813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,256,0.03348355491956075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,128,0.03369333346684774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,128,0.03913777735498216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,64,0.03031377659903632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,64,0.039926220973332725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,16384,32,0.030986666679382324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,16384,32,0.04007110993067423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,16384,128,0.03272622161441379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,65536,4.579387664794922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,65536,2.5864791870117188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,16384,0.5918986532423232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,16384,1.201738675435384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,16384,0.6300622092352973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,65536,2.798203574286567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,12288,0.8303164376152886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,12288,0.49447022544013125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,10240,0.37766133414374453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,10240,0.7293928994072808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,10240,0.41129599677191836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,12288,0.4383315510219998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,8192,0.5446809132893881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,8192,0.3293644322289361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,8192,0.2979831165737576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,7168,0.48737510045369464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,7168,0.28520533773634166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,7168,0.2608666684892443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,6144,0.39968355496724445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,6144,0.24569866392347547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,6144,0.22528799374898276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,5120,0.3371253278520372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,5120,0.20527733696831596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,5120,0.18869600031110975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,4096,0.27048977216084796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,4096,0.16212977303398982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,3584,0.14279022481706408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,3584,0.23994665675693086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,3584,0.1423271099726359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,4096,0.15357777807447645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,3072,0.20666310522291395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,3072,0.12194400363498265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,2560,0.10417422321107651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,2560,0.17461600568559435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,2560,0.10395199722713894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,3072,0.11777510907914902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,2048,0.14023555649651423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,2048,0.08328355683220758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,2048,0.08461244238747491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,1536,0.10871555407842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,1536,0.06375377708011203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,1536,0.06708621978759766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,1024,0.07657688856124878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,1024,0.04917066627078586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,1024,0.05073422193527222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,768,0.0610062215063307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,768,0.04292800029118856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,768,0.04320711228582594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,512,0.04541866646872627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,512,0.03895555602179633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,512,0.03296533226966858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,256,0.030478222502602473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,256,0.033725334538353816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,256,0.027238221632109746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,128,0.026239110363854304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,128,0.03174577818976508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,64,0.02407022151682112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,64,0.03281777766015794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,12288,32,0.025095999240875244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,12288,32,0.03279111120435927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,12288,128,0.02588444451491038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,65536,2.0252827538384333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,65536,3.918553670247396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,65536,2.282491472032335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,16384,0.986936887105306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,16384,0.5360462400648329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,16384,0.49494043986002606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,12288,0.7063102192348905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,12288,0.4160462220509847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,12288,0.3701359960767958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,10240,0.574019538031684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,10240,0.3474675549401178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,10240,0.31199555926852757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,8192,0.45520888434516055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,8192,0.2690737777286106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,8192,0.2504444387223986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,7168,0.3950666586558024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,7168,0.2414862314860026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,7168,0.2213644451565213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,6144,0.3462079895867242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,6144,0.20448088645935059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,6144,0.19003555509779188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,5120,0.28509333398607045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,5120,0.1734480063120524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,5120,0.1604879962073432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,4096,0.2311840057373047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,4096,0.1380231115553114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,4096,0.12986311647627088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,3584,0.20279467105865479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,3584,0.12225511338975693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,3584,0.11590577496422662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,3072,0.17631022135416666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,3072,0.10474222236209446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,3072,0.10000622272491455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,2560,0.14837600125206843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,2560,0.08764888842900594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,2560,0.08678311109542847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,2048,0.12156532870398627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,2048,0.0695635543929206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,2048,0.07189689079920451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,1536,0.09399111403359307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,1536,0.055431998438305326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,1024,0.04407022065586514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,1024,0.0661662220954895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,1024,0.04236533244450887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,1536,0.057987557517157666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,768,0.05303111010127597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,768,0.03723377651638455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,768,0.03757066528002421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,512,0.04002133342954848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,512,0.033788445923063494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,512,0.02908266584078471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,256,0.02828533450762431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,256,0.029487109846538965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,256,0.02375999920898014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,10240,128,0.022188444932301838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,128,0.022279111875428095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,128,0.02798844377199809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,64,0.019692444139056735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,64,0.02867377797762553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,10240,32,0.02018666598531935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,10240,32,0.028439998626708984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,65536,3.1766274770100913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,65536,1.8391458723280165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,65536,1.6766737831963434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,16384,0.7297466595967611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,16384,0.41748799218071836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,16384,0.4013208813137478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,12288,0.5346230930752224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,12288,0.3320008913675944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,12288,0.30154667960272896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,10240,0.4488933351304796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,10240,0.28038133515252006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,8192,0.2354257901509603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,8192,0.36890043152703184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,8192,0.21183199352688262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,10240,0.25439288881089955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,7168,0.3122977680630154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,7168,0.19268355104658338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,7168,0.17994844913482666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,6144,0.26827200253804523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,6144,0.16685333516862658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,6144,0.15593155225118002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,5120,0.22545599937438965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,5120,0.14006133874257407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,5120,0.13101066483391657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,4096,0.18121600151062012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,4096,0.11379644605848525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,4096,0.1071831120385064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,3584,0.16047999593946669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,3584,0.09722222222222222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,3584,0.09487555424372356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,3072,0.13850311438242593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,3072,0.08525955677032471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,3072,0.08351644542482163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,2560,0.11744177341461182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,2560,0.071070220735338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,2560,0.07144533263312446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,2048,0.09530044264263576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,2048,0.057405332724253334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,2048,0.060025778081681996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,1536,0.07385244634416369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,1536,0.04597333404752943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,1536,0.04825244347254435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,1024,0.052729778819613986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,1024,0.03583555420239767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,1024,0.0366195539633433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,768,0.042315555943383105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,768,0.03147466646300422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,768,0.031238221459918555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,512,0.03211733367707994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,512,0.028385778268178303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,512,0.024683554967244465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,256,0.022379555635982092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,256,0.024889777104059856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,256,0.02014133334159851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,128,0.017887110511461895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,128,0.02386222283045451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,64,0.01497866710027059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,64,0.024302222662501868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,8192,32,0.015787555111779105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,8192,32,0.02407199972205692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,8192,128,0.018350222044520907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,65536,2.823724534776476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,65536,1.6199359893798828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,65536,1.6350133683946397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,16384,0.6521040068732368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,16384,0.3924684524536133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,16384,0.4008791181776259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,12288,0.4671529134114583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,12288,0.3023306793636746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,10240,0.2772257857852512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,10240,0.4023279878828261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,10240,0.2473564412858751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,12288,0.3013288974761963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,8192,0.31035733222961426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,8192,0.19351643986172148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,8192,0.2053928905063205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,7168,0.27399733331468373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,7168,0.1722302171919081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,7168,0.17999555004967582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,6144,0.23604888386196562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,6144,0.14995288848876953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,6144,0.15610399511125353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,5120,0.19785600238376191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,5120,0.12600533167521158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,5120,0.1315279934141371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,4096,0.15901688734690347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,4096,0.10062399837705825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,3584,0.09592888752619426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,3584,0.13956977261437312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,3584,0.08826755815082127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,4096,0.10766400231255426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,3072,0.12072000238630508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,3072,0.07574222485224406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,2560,0.07288355297512479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,2560,0.10097955332862006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,2560,0.06449066930347018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,3072,0.08396622207429673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,2048,0.08072621954811944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,2048,0.0520995557308197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,2048,0.06089244286219279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,1536,0.06197244591183133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,1536,0.041965332296159535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,1536,0.048956443866093956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,1024,0.04328533344798618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,1024,0.032631112469567194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,1024,0.03928800092803107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,768,0.03458222084575229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,768,0.02896711230278015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,768,0.03234133455488417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,512,0.025271111064487036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,512,0.0259226659933726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,256,0.02106222179200914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,256,0.017737777696715463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,256,0.02260533306333754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,512,0.026503110925356548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,128,0.014615111880832247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,128,0.02095555596881443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,64,0.013728888498412238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,64,0.021103999680942957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,7168,32,0.014383110735151501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,7168,32,0.02079288827048408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,7168,128,0.018802666001849704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,65536,2.350203620062934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,65536,1.418028407626682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,65536,1.2861404418945312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,16384,0.5592640240987142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,16384,0.34569689962599015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,16384,0.3087831073337131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,12288,0.3951626618703206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,12288,0.25702667236328125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,12288,0.23482844564649794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,10240,0.3321688969930013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,10240,0.21810400485992432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,10240,0.19630755318535698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,8192,0.2660568820105659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,8192,0.16682755947113037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,8192,0.1594746642642551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,7168,0.23475644323560926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,7168,0.14916711383395725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,7168,0.14006488853030735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,6144,0.20155289438035753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,6144,0.13135822614034018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,6144,0.12190311484866673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,5120,0.16976355181799996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,5120,0.1103057795100742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,5120,0.10250488917032878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,4096,0.13704000578986272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,4096,0.08667555782530044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,4096,0.0848204427295261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,3584,0.12138133578830296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,3584,0.07694666915469699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,3584,0.07517511314815946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,3072,0.10515644152959187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,3072,0.06642133659786649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,3072,0.06599289178848267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,2560,0.0890008873409695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,2560,0.05678577555550469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,2560,0.056856890519460045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,2048,0.07256266805860731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,2048,0.04608088731765747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,2048,0.04789244466357761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,1536,0.0565777752134535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,1536,0.03714311122894287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,1536,0.03885066509246826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,1024,0.040757334894604154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,1024,0.029209779368506536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,1024,0.029849777619043987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,768,0.03325600094265408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,768,0.02588622272014618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,768,0.02574133376280467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,512,0.025772444076008264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,512,0.02316177719169193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,512,0.02088177700837453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,256,0.018260444204012554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,256,0.01998133295112186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,256,0.016917333006858826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,128,0.01421866648726993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,128,0.018453333112928603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,64,0.012353777885437012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,64,0.018087110585636564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,6144,32,0.012476444244384766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,6144,32,0.017670222454600863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,6144,128,0.015455999308162265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,65536,1.8700586954752605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,65536,1.2049119737413194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,65536,1.3141742282443576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,16384,0.4507146411471897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,16384,0.28966132799784344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,16384,0.30686577161153156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,12288,0.33390400144788956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,12288,0.23044888178507486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,12288,0.21954933802286783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,10240,0.28142134348551434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,10240,0.18448711766137016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,10240,0.19376444816589355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,8192,0.2238044473859999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,8192,0.14282222588857016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,8192,0.1579751173655192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,7168,0.19694666067759195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,7168,0.12707377804650202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,7168,0.13964533805847168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,6144,0.16916355821821424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,6144,0.11318400171067979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,6144,0.12108622656928168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,5120,0.14225867059495714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,5120,0.09427289168039958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,5120,0.10186755657196045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,4096,0.11468888653649224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,4096,0.07377866903940837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,4096,0.08401778009202744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,3584,0.10145066844092475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,3584,0.06501155429416232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,3584,0.075162669022878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,3072,0.08759822448094685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,3072,0.05713866816626655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,3072,0.06619822316699557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,2560,0.07447555330064562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,2560,0.04912799927923414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,2560,0.057328886455959745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,2048,0.06052177482181125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,2048,0.04021866785155402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,2048,0.048496001296573214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,1536,0.047206223011016846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,1536,0.03269777695337931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,1536,0.0396053327454461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,1024,0.033843557039896645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,1024,0.025864000121752422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,1024,0.03198933270242479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,768,0.027038221557935078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,768,0.022870222727457683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,768,0.026177777184380427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,512,0.020376889242066275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,512,0.019887111253208585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,512,0.022397332721286353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,256,0.01427288850148519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,256,0.017044444878896076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,128,0.015421332584487068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,128,0.010926222635640038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,128,0.015430221954981485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,64,0.009187555147541894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,64,0.015200000670221118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,5120,32,0.009351111120647853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,5120,32,0.015132443772421943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,5120,256,0.016960889101028442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,65536,1.4661137262980144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,65536,0.8995226754082574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,65536,0.9862035115559896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,16384,0.3635866641998291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,16384,0.2391288810306125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,16384,0.2197697824902005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,12288,0.26265154944525826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,12288,0.17914311091105142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,12288,0.1659875578350491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,10240,0.21829156080881754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,10240,0.1529795593685574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,10240,0.14056266678704157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,8192,0.17652266555362275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,8192,0.11838400363922119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,8192,0.11354488796657985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,7168,0.15441777971055773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,7168,0.10524355702930027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,7168,0.10109955734676784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,6144,0.13367644945780435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,6144,0.0874977774090237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,6144,0.09243288967344497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,5120,0.11207111676534016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,5120,0.07686489158206515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,5120,0.07439466979768541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,4096,0.09047555261188084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,4096,0.061237335205078125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,4096,0.06222933530807495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,3584,0.0551457769340939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,3584,0.054690665668911405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,3584,0.08027911186218262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,3072,0.0697466664844089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,3072,0.04782311121622721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,3072,0.04840888910823398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,2560,0.05874133110046387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,2560,0.04119377666049533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,2560,0.04227644536230299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,2048,0.04831111099984911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,2048,0.03403110967742072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,2048,0.03558844327926636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,1536,0.037089778317345515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,1536,0.027786665492587622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,1536,0.029192888074451025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,1024,0.026497777965333726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,1024,0.022272888157102797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,1024,0.023220444718996685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,768,0.02160444524553087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,768,0.019896000623703003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,768,0.0199306673473782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,512,0.01664888858795166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,512,0.0173653331067827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,512,0.016833777228991192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,256,0.0118222220076455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,256,0.015063999427689446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,256,0.013631111217869652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,128,0.00907199995385276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,128,0.01296266665061315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,64,0.007889777421951294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,64,0.013391111459996967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,4096,32,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,4096,32,0.013420444395807056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,4096,128,0.012679110798570844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,65536,1.433295143975152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,65536,0.9796142578125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,65536,0.8834400177001953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,16384,0.35030044449700254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,16384,0.23307821485731336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,16384,0.21926221582624647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,12288,0.2494586573706733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,12288,0.17872267299228242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,10240,0.1480648914972941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,10240,0.21678133805592856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,10240,0.15138044622209337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,12288,0.16649511125352648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,8192,0.16724888483683267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,8192,0.11777155929141575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,8192,0.11479822794596355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,7168,0.1473671065436469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,7168,0.10327466991212632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,7168,0.1007680031988356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,6144,0.12667022811041936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,6144,0.09205600288179185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,6144,0.08867110808690389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,5120,0.1068248881234063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,5120,0.07636711332533094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,5120,0.07472888628641765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,4096,0.08619644244511922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,4096,0.060566220018598765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,4096,0.06236444579230415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,3584,0.0764604475763109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,3584,0.053295112318462796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,3584,0.05585244629118177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,3072,0.06628800100750394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,3072,0.04703111118740506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,3072,0.0491537782880995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,2560,0.05623200204637316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,2560,0.04051999913321601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,2560,0.043028444051742554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,2048,0.046022223101721875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,2048,0.03327644533581204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,2048,0.03631022241380479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,1536,0.03566844595803155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,1536,0.02757511039574941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,1536,0.029907554388046265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,1024,0.025861332813898723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,1024,0.022297778063350256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,1024,0.024736000431908503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,768,0.021149333980348375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,768,0.019691555036438834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,768,0.0207342223988639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,512,0.016768889294730294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,512,0.017326222525702584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,512,0.01772888832622104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,256,0.011567110816637674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,256,0.01460977726512485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,128,0.012818666795889536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,128,0.008750222623348236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,128,0.012596444123321109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,64,0.007528000407748752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,64,0.013018666870064206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3584,32,0.007801777786678738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3584,32,0.013106666505336761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3584,256,0.013810666898886362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,65536,1.115093337164985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,65536,0.8083768950568305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,65536,0.7025430997212728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,16384,0.2738017770979139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,16384,0.18715288903978136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,16384,0.17395288414425322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,12288,0.2041840023464627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,12288,0.14517688751220703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,12288,0.13303200403849283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,10240,0.1716453366809421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,10240,0.12196000417073567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,10240,0.1111946635776096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,8192,0.1372728877597385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,8192,0.09417421950234307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,8192,0.09134933021333481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,7168,0.12125689453548855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,7168,0.08391111426883274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,7168,0.08045866754319933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,6144,0.10455911027060615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,6144,0.07397600015004475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,6144,0.07088977760738797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,5120,0.08791200319925944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,5120,0.06279555294248793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,5120,0.05990933047400581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,4096,0.07069333394368489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,4096,0.04970844586690267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,4096,0.04966577887535095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,3584,0.0626879996723599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,3584,0.04445599847369724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,3584,0.044418666097852916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,3072,0.05425688955518934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,3072,0.039112889104419284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,3072,0.03920800156063504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,2560,0.045863999260796436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,2560,0.03369955552948846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,2560,0.034064001507229276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,2048,0.03731733229425218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,2048,0.027822222974565294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,2048,0.029280887709723577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,1536,0.028951999213960435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,1536,0.023473777704768713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,1536,0.024528889192475214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,1024,0.02077155477470822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,1024,0.0192248887485928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,1024,0.019799111617936026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,768,0.01699822147687276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,768,0.017087111870447796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,768,0.01739644507567088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,512,0.013371555341614617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,512,0.015043555034531487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,512,0.01495733360449473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,256,0.009197333620654212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,256,0.012700444294346703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,256,0.012337778177526263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,128,0.007319110963079665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,128,0.011213333242469363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,64,0.0063662222690052455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,64,0.011600000162919363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,3072,32,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,3072,32,0.011608888705571493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,3072,128,0.010891555911964841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,65536,0.8617946836683484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,65536,0.8179004457261828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,65536,0.606279108259413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,16384,0.2168817785051134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,16384,0.18626133600870767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,16384,0.1553528838687473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,12288,0.16469244162241617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,12288,0.14046311378479004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,10240,0.10157510969373916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,10240,0.13828444480895996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,10240,0.11933688322703044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,12288,0.11789600054423015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,8192,0.11064800288942124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,8192,0.0914471083217197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,8192,0.08152800136142306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,7168,0.09788178073035346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,7168,0.08161777920193143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,7168,0.07155466741985746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,6144,0.08409421973758274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,6144,0.07155911127726237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,6144,0.0630177789264255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,5120,0.07109333409203424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,5120,0.06057066387600369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,5120,0.05354311068852743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,4096,0.05754044320848253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,4096,0.047651555803087026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,3584,0.03959111041492886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,3584,0.05087111062473721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,3584,0.04275822308328417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,4096,0.044403556320402354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,3072,0.04442044430308872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,3072,0.037724445263544716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,2560,0.03088888857099745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,2560,0.03771377934349908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,2560,0.0324942237801022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,3072,0.03513600097762214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,2048,0.030761778354644775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,2048,0.027323555615213182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,2048,0.02657866643534766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,1536,0.024578667349285547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,1536,0.02310133311483595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,1536,0.0221368885702557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,1024,0.017986666825082567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,1024,0.01906222270594703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,1024,0.017861333158281114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,768,0.014947555131382413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,768,0.016910221841600206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,768,0.015592000550693937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,512,0.01185955521133211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,512,0.014703111516104804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,256,0.010977778169843884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,256,0.008349333372380998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,256,0.012106666962305704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,512,0.013293333351612091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,128,0.006893333461549547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,128,0.01090133355723487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,64,0.006071110980378256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,64,0.011298666397730509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2560,128,0.009873777627944946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2560,32,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2560,32,0.011524444652928246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,65536,0.6898595492045084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,65536,0.6176648669772679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,65536,0.513662232293023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,16384,0.17439466052585176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,16384,0.1413831048541599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,16384,0.13230844338734946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,12288,0.13120088312360975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,12288,0.11089333560731675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,12288,0.10068888796700372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,10240,0.1098124451107449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,10240,0.09486133522457546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,10240,0.08575999736785889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,8192,0.08854844172795613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,8192,0.07038844294018216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,8192,0.06971377796596952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,7168,0.0625822212960985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,7168,0.07795555724038018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,7168,0.06346488661236234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,6144,0.06752000252405803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,6144,0.05671733617782593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,5120,0.04688177837265862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,5120,0.0570160018073188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,5120,0.04853244291411506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,6144,0.054564442899492055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,4096,0.04617244336340162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,4096,0.03881866733233134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,4096,0.0387644436624315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,3584,0.0410426656405131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,3584,0.033760888708962336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,3584,0.03475288881195916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,3072,0.03559733430544535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,3072,0.029880000485314265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,3072,0.03084355592727661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,2560,0.030441777573691473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,2560,0.026216889421145122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,2560,0.027331555883089702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,2048,0.024999111890792847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,2048,0.021944888763957556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,1536,0.020053333706325956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,1536,0.020032000210550096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,1536,0.018923555811246235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,2048,0.02363377809524536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,1024,0.015040889382362366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,1024,0.015584889385435315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,768,0.014404444230927361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,768,0.012620444099108378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,768,0.014167999227841696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,1024,0.016390222642156813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,512,0.00999022192425198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,512,0.012120000190205045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,512,0.012627555264366997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,256,0.00721688899728987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,256,0.010229333407349056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,256,0.010755555497275459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,128,0.006172444257471297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,128,0.00960000024901496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,64,0.005394666559166378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,64,0.009938666390048133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,2048,32,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,2048,32,0.009893333746327294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,2048,128,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,65536,0.5336435635884603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,65536,0.5951075553894043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,65536,0.41990844408671063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,16384,0.1341866652170817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,16384,0.14278666178385416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,12288,0.08536711004045273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,12288,0.10193955236011082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,12288,0.10357777939902411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,16384,0.10926844676335652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,10240,0.08544177479214138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,10240,0.08766400151782566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,10240,0.07037244240442912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,8192,0.0687191088994344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,8192,0.06889777713351779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,8192,0.05806489123238457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,7168,0.06097333298789131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,7168,0.06147022379769219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,7168,0.051522665553622775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,6144,0.052384889788097806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,6144,0.05505422088834974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,6144,0.04571733209821913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,5120,0.044360889328850635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,5120,0.046866665283838906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,5120,0.03877066572507223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,4096,0.035888890425364174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,4096,0.036890665690104164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,4096,0.032280001375410296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,3584,0.031940443648232356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,3584,0.032919999625947736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,3584,0.029112001260121662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,3072,0.027656000521447923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,3072,0.02942311101489597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,3072,0.0262106673585044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,2560,0.023822221491071913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,2560,0.026025777061780293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,2560,0.023279110590616863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,2048,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,2048,0.019854222734769184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,2048,0.022083555658658344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,1536,0.015965332587560017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,1536,0.01889955500761668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,1536,0.017232000827789307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,1024,0.01203733351495531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,1024,0.015469332536061605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,1024,0.01441333360142178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,768,0.010256888965765635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,768,0.01425244410832723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,768,0.012885332935386233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,512,0.007939555578761632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,512,0.011857777833938599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,512,0.011516444385051727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,256,0.006039999839332368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,256,0.010207110808955299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,128,0.008741333252853816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,128,0.005111999809741974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,128,0.009547555612193214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,64,0.004592000196377437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,64,0.009734222458468543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1536,32,0.004790222065316306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1536,32,0.009763555394278632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1536,256,0.009783999787436591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,65536,0.35980799463060165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,65536,0.4229288895924886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,65536,0.3540986643897162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,16384,0.0903093351258172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,16384,0.0987973345650567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,16384,0.08955466747283936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,12288,0.06859378019968669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,12288,0.07203644514083862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,12288,0.0696444445186191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,10240,0.05739733245637682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,10240,0.06258310874303182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,10240,0.05870133638381958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,8192,0.046801779005262584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,8192,0.04962133367856344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,8192,0.0481644438372718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,7168,0.041167110204696655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,7168,0.04466133316357931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,7168,0.04331022169854906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,6144,0.035826666487587824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,6144,0.039721777041753135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,6144,0.03782400157716539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,5120,0.030079109801186457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,5120,0.034865776697794594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,5120,0.03226222263442145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,4096,0.024708444873491924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,4096,0.02703199949529436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,4096,0.026769777139027912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,3584,0.021964444054497614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,3584,0.024719999896155462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,3584,0.02466222147146861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,3072,0.01936444474591149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,3072,0.022514666120211285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,3072,0.022110222114457026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,2560,0.016615110966894362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,2560,0.020278222031063504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,2560,0.019671999745898776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,2048,0.014158222410413953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,2048,0.017807111144065857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,2048,0.017135999268955655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,1536,0.011519110865063138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,1536,0.015560888581805758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,1024,0.012466666599114736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,1024,0.009069333473841349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,1024,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,768,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,768,0.007782222496138678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,768,0.011985777980751462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,1536,0.01479111115137736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,512,0.00619555554456181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,512,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,512,0.00981600003110038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,256,0.004998222407367495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,256,0.008776889079146916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,256,0.007726222276687622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,128,0.004283555679851108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,128,0.008227555288208855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,64,0.003952888978852166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,64,0.008402666283978356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,1024,32,0.004139555411206352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,1024,32,0.008435555630260045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,1024,128,0.0069857777820693115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,65536,0.27748266855875653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,65536,0.3910115559895833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,65536,0.3372773329416911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,16384,0.07224177651935153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,16384,0.08790044652091132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,16384,0.08820177449120416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,12288,0.055418666866090566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,12288,0.06636533472273085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,10240,0.0583475563261244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,10240,0.04579822222391764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,10240,0.05569066603978475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,12288,0.06776977909935845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,8192,0.037046220567491316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,8192,0.045104887750413686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,8192,0.04744266801410251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,7168,0.03269422385427687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,7168,0.040892445378833346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,7168,0.0425217780801985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,6144,0.02868177824550205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,6144,0.0370124446021186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,6144,0.03677422139379714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,5120,0.02442933287885454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,5120,0.03189244535234239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,5120,0.031503111124038696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,4096,0.01982399986849891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,4096,0.02513955533504486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,4096,0.026280888252788123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,3584,0.01778133379088508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,3584,0.022778666681713525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,3584,0.02411288850837284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,3072,0.017329777280489605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,3072,0.02091022166940901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,3072,0.021552888883484736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,2560,0.01425333321094513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,2560,0.01901511185699039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,2560,0.019298666053348117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,2048,0.012504888905419244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,2048,0.01661422186427646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,2048,0.016777777009540133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,1536,0.009775111244784461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,1536,0.014874666929244995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,1536,0.014475555883513557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,1024,0.007752888732486301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,1024,0.01275288893116845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,1024,0.012221333053376941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,768,0.006763555523422029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,768,0.011624000138706632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,768,0.010277333358923594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,512,0.005390222287840313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,512,0.009561777942710454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,512,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,256,0.004547555413511065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,256,0.008635555704434713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,128,0.0063573333124319715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,128,0.003976888954639435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,128,0.008051555189821456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,64,0.003628444340493944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,64,0.00817599975400501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,768,32,0.0038195554580953387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,768,32,0.008168000313970778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,512,0.012072000238630505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,65536,0.19186045063866508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,65536,0.36744001176622176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,768,256,0.0068915556702348925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,16384,0.053780443138546415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,16384,0.08549066384633382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,65536,0.29155556360880536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,12288,0.048617776897218495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,12288,0.06193155712551541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,16384,0.07515377468532987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,10240,0.043154666821161904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,10240,0.05415466758939955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,12288,0.060271998246510826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,8192,0.032207111517588295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,8192,0.04371910956170824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,512,0.008187555604510838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,7168,0.02882666720284356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,7168,0.03987733191914029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,10240,0.051870223548677236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,6144,0.025362667110231187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,6144,0.035614222288131714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,5120,0.027983112467659846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,5120,0.022429333792792425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,5120,0.031517333454555936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,8192,0.04131022095680237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,4096,0.01808177762561374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,4096,0.02472266720400916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,7168,0.036970665057500206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,3584,0.015593777100245157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,3584,0.022485334012243483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,6144,0.03287555442916022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,3072,0.015013333823945789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,3072,0.020627554919984605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,2560,0.01718488832314809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,2560,0.012831110921170978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,2560,0.018533334136009216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,4096,0.02380977736579047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,2048,0.010652444428867765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,2048,0.016384000579516094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,3584,0.0214711113108529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,1536,0.008783111141787635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,1536,0.014530667000346713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,2048,0.01500266624821557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,3072,0.019313777486483257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,1024,0.006811555474996567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,1024,0.012324444121784635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,768,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,768,0.005727111051479976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,768,0.011201777391963534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,512,0.004774222357405556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,512,0.009570666485362584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,256,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,256,0.003992889076471329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,256,0.008527999950779809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,1024,0.010760000182522668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,128,0.0035982223020659555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,128,0.007883555359310573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,1536,0.012968888713253869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,64,0.0032755554550223877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,512,32,0.0034231110993358824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,64,0.007995555798212687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,512,32,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,65536,0.11475200123257107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,65536,0.36649423175387913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,512,0.007341333561473423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,16384,0.038731555143992104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,16384,0.08548888895246719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,512,128,0.0058444444504049085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,12288,0.0311173333062066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,12288,0.06206399864620633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,65536,0.2641635470920139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,10240,0.030321776866912842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,10240,0.052856889035966664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,16384,0.0681697792477078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,8192,0.01905333333545261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,8192,0.0429111123085022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,12288,0.05299555593066745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,7168,0.020220445262061227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,7168,0.03898488812976413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,10240,0.04618666569391886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,6144,0.018386666973431904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,6144,0.034676445855034724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,5120,0.02497244377930959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,5120,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,5120,0.03045066528850132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,8192,0.037603557109832764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,4096,0.012515555653307172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,7168,0.03329244587156508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,4096,0.023955555425749883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,3584,0.011343110766675739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,3584,0.0221377776728736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,6144,0.02920177910063002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,3072,0.010320000350475311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,3072,0.020297777321603563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,4096,0.021078222327762183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,2560,0.008824889030721452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,2560,0.018459555175569322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,2048,0.013473778135246701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,2048,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,2048,0.016235555211702984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,3584,0.01921422282854716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,1536,0.006776888751321369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,1536,0.014463111758232117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,3072,0.017297777864668105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,1024,0.004795555439260271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,1024,0.012140444583363004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,2560,0.015403555499182807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,768,0.004337777694066365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,768,0.010398222340477837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,1536,0.011379555695586734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,512,0.003871110992299186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,512,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,256,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,256,0.0034426665968365143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,256,0.008345777789751688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,1024,0.009173333644866943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,128,0.003156444471743372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,128,0.00777422222826216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,64,0.00300799993177255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,64,0.007872000336647034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,256,32,0.003118222165438864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,256,32,0.007880888879299164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,768,0.007623111208279927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,65536,0.22513156467013887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,65536,0.10404000017378066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,65536,0.36578578419155544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,16384,0.03244533472590976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,16384,0.08623822530110677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,512,0.006503110958470239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,12288,0.025711110896534387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,12288,0.06232711341645983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,256,128,0.0053546664615472155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,10240,0.022469333476490442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,10240,0.05312444435225593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,8192,0.03281777766015794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,8192,0.018961777289708454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,8192,0.0428711109691196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,7168,0.02908888790342543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,7168,0.017198221551047433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,7168,0.03887110948562622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,16384,0.060106668207380504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,6144,0.014659555421935188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,6144,0.03436888919936286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,5120,0.02164711058139801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,5120,0.013179555535316467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,5120,0.030232889784706965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,12288,0.04673244555791219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,4096,0.01072622256146537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,4096,0.024314666787783306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,10240,0.0398613346947564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,3584,0.009515555368529426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,3584,0.022368000613318548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,6144,0.025416889124446448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,3072,0.00909511082702213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,3072,0.02050044470363193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,2560,0.013635555903116861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,2560,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,2560,0.018383999665578205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,2048,0.012177777787049612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,2048,0.007152888509962294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,2048,0.016189333465364244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,1536,0.010384000009960597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,1536,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,4096,0.01834222177664439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,1536,0.014503111441930136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,1024,0.00480266660451889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,1024,0.012168889244397482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,3584,0.016869333055284288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,768,0.004135999828577042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,768,0.010663111176755695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,512,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,512,0.0036551112102137674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,512,0.00888533310757743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,3072,0.015231110983424716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,256,0.0033155555526415506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,256,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,1024,0.008086221913496653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,768,0.007249777515729268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,128,0.007814221911960179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,128,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,64,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,128,32,0.0029973333908451926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,64,0.007800000409285228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,128,32,0.007849777738253275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,65536,0.09857066472371419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,16384,0.029030223687489826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,16384,0.0857422219382392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,12288,0.023637332850032385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,65536,0.365847110748291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,12288,0.06202399730682373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,10240,0.02307111190425025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,8192,0.017321777012613084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,10240,0.05270222160551283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,8192,0.04279911186960009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,7168,0.015896000795894198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,7168,0.038797332180870905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,6144,0.014459555347760519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,6144,0.0344124436378479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,5120,0.012645333177513547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,5120,0.030254221624798242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,4096,0.011215999722480774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,4096,0.024284443921513025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,3584,0.009279111193286048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,3584,0.02234933276971181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,3072,0.008231999973456064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,3072,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,2560,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,2560,0.018181333939234417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,2048,0.006808888994985157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,2048,0.01610488858487871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,1536,0.005495999836259418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,1024,0.0041573333243529005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,1536,0.014260444376203748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,1024,0.01184977756606208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,768,0.0038284444146686127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,768,0.009857777920034196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,512,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,256,0.003140444349911478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,512,0.008784000244405534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,128,0.0028604444944196274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,256,0.008192888564533657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,64,0.0027457777824666765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,128,0.007668444679843054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,64,32,0.0027804445061418745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,64,0.00776533368561003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,64,32,0.007748444047239091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,65536,0.09948177470101251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,65536,0.364208009507921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,16384,0.027907556957668726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,12288,0.022699554761250813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,12288,0.06196000178654989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,16384,0.08559466732872857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,10240,0.019903111788961623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,10240,0.052467554807662964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,8192,0.016550223032633465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,7168,0.015024888846609326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,8192,0.0425217780801985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,7168,0.03852533300717672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,6144,0.013765333427323235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,6144,0.034048000971476235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,5120,0.012392889294359418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,4096,0.010862222148312463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,5120,0.03010400136311849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,4096,0.024142222272025213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,3584,0.010271111296282874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,3584,0.022255111071798537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,3072,0.009569777382744683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,3072,0.02017511096265581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,2560,0.0084906667470932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,2048,0.006671999891599019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,2560,0.0181031111213896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,2048,0.016172443826993305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,1536,0.0057653333577844836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,1024,0.004456888884305954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,1536,0.01423822177780999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,1024,0.01187111106183794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,768,0.003830222205983268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,768,0.009766221874290043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,512,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,512,0.008793777889675563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,256,0.0031617778456873367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,256,0.008177777959240807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,128,0.0029120000286234748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,128,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,64,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,64,0.007734222544564142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2048,32,32,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2048,32,32,0.007700444095664554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,16384,3.413886176215278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,16384,1.641183959113227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,16384,1.6558586756388347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,12288,2.5415022108289933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,12288,1.0890639623006184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,12288,1.179250717163086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,10240,2.1174123552110458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,10240,0.9721715715196398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,10240,0.9272666507297092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,8192,1.7232667075263128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,8192,0.7618417739868164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,8192,0.7806622187296549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,7168,1.5033004548814561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,7168,0.6708195474412707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,7168,0.6821617550320096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,6144,1.278429349263509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,6144,0.5544488694932725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,6144,0.578096866607666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,5120,1.037965350680881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,5120,0.46197954813639325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,4096,0.3924391004774306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,5120,0.4801040225558811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,4096,0.3640151023864746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,4096,0.8150080045064291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,3584,0.7308497958713107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,3584,0.32896712091234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,3584,0.34030667940775555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,3072,0.5681537522210015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,3072,0.2785511016845703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,3072,0.2966870996687147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,2560,0.4841173489888509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,2560,0.2445697784423828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,2560,0.25104355812072754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,2048,0.20853600237104628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,2048,0.38991554578145343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,2048,0.19123378064897326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,1536,0.3013555473751492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,1536,0.15187288655175105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,1024,0.1264879968431261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,1024,0.2116373380025228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,1024,0.1259795559777154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,768,0.10479022396935357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,768,0.16676178243425158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,768,0.10379022359848022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,1536,0.16276444329155818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,512,0.11752177609337701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,512,0.08802844418419732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,512,0.07758400175306532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,256,0.07194488578372531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,256,0.0752355522579617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,256,0.06335910823610094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,128,0.06334311432308622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,128,0.07172266642252605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,64,0.05817511346605089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,64,0.07309333483378093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,65536,32,0.057895110713111035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,65536,32,0.07394666804207696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,65536,128,0.06277333365546332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,65536,3.233088811238607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,65536,1.6451209386189778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,16384,0.39721067746480304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,16384,0.7452426486545138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,65536,1.7267430623372395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,16384,0.3828293217553033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,12288,0.2928231027391222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,12288,0.5588471094767252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,12288,0.2825831042395698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,10240,0.2872515519460042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,10240,0.44871287875705296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,10240,0.24224445554945204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,8192,0.35811556710137266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,8192,0.19814577367570665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,7168,0.1795626613828871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,7168,0.3208186626434326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,7168,0.17171288861168754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,6144,0.15459911028544107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,6144,0.2951280011071099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,6144,0.17022132873535156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,8192,0.19497333632575142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,5120,0.22419733471340605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,5120,0.12200178040398492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,5120,0.12692177295684814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,4096,0.18277866310543486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,4096,0.10044178035524155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,3584,0.09474844402737087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,3584,0.16174311108059353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,3584,0.08713066577911377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,4096,0.10323822498321533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,3072,0.1385671165254381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,3072,0.07487822241253324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,3072,0.08067111174265544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,2560,0.11818933486938477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,2560,0.0646266672346327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,2560,0.06896799802780151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,2048,0.09586666689978705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,2048,0.05331200030114916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,2048,0.05808266666200426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,1536,0.07460977633794148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,1536,0.043268445465299815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,1024,0.03596977723969354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,1024,0.053343110614352755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,1536,0.046690666013293795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,1024,0.03466577662361993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,768,0.04273066586918301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,768,0.030805332793129817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,768,0.031247112486097548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,512,0.024940444363488093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,512,0.0325564444065094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,512,0.027595554788907368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,256,0.02281422250800663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,256,0.023815110325813293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,256,0.02012800011369917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,128,0.018352889352374606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,16384,128,0.018365333477656048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,128,0.02276977731121911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,64,0.015094222293959724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,64,0.02352977792421977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,16384,32,0.01553600033124288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,16384,32,0.02307555576165517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,65536,1.1726115544637044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,65536,2.4421564737955728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,65536,1.2613804075453017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,16384,0.5887395540873209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,16384,0.33297954665289986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,12288,0.22636000315348306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,12288,0.40517155329386395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,12288,0.24268621868557402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,16384,0.29091911845737034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,10240,0.3376399940914578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,10240,0.20233955648210314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,10240,0.18659733401404488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,8192,0.2677324348025852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,8192,0.14659910731845432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,7168,0.13363111019134521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,7168,0.23601065741644967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,7168,0.13142666551801893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,8192,0.1495866643057929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,6144,0.20110400517781576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,6144,0.11232088671790229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,5120,0.0981466637717353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,5120,0.16951110627916124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,5120,0.09408266676796807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,4096,0.08223822381761339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,4096,0.13754399617513022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,4096,0.07959910896089342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,6144,0.11535199483235677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,3584,0.12195377879672581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,3584,0.0678079989221361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,3072,0.06315288941065471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,3072,0.10621777507993911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,3072,0.05920355849795871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,3584,0.07138399945365058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,2560,0.08894399801890056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,2560,0.05063022176424662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,2048,0.045796444018681846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,2048,0.07291111018922594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,2048,0.041962666643990405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,2560,0.05441688828998142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,1536,0.05695822503831652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,1536,0.034394668208228216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,1536,0.03767555620935228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,1024,0.040844443771574236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,1024,0.027951998843087092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,768,0.02548888822396596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,768,0.03404711021317376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,768,0.025156444973415796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,1024,0.02930844492382473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,512,0.025822222232818604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,512,0.021987555755509272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,256,0.016774222254753113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,256,0.018301332990328472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,256,0.018637332651350234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,512,0.02082577844460805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,128,0.014314666390419006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,128,0.01683822274208069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,64,0.01275022245115704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,64,0.016376889414257474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,12288,32,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,12288,32,0.01678400072786543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,65536,1.2183546490139432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,12288,128,0.01532977819442749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,65536,2.0578471289740667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,65536,1.0642400317721896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,16384,0.49310933219061953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,16384,0.2544204394022624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,12288,0.22372712029351127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,12288,0.3365066581302219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,12288,0.20000089539421928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,16384,0.28602578904893666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,10240,0.2806657685173882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,10240,0.1621448861228095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,10240,0.1830159955554538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,8192,0.2265413337283664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,8192,0.12325867017110188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,8192,0.14707733525170222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,7168,0.19732177257537842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,7168,0.11112266116672093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,7168,0.13095200061798096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,6144,0.17164177364773223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,6144,0.09416533178753322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,5120,0.09654221932093303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,5120,0.14426488346523708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,5120,0.08019910918341742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,4096,0.08040355311499701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,4096,0.11660088433159722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,4096,0.06506133079528809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,6144,0.11310222413804795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,3584,0.10211555825339423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,3584,0.05739733245637682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,3072,0.06244444184833103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,3072,0.08839555581410725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,3072,0.05045688814587063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,2560,0.0547653337319692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,2560,0.07487999730639987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,2560,0.043840001026789345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,3584,0.07175733645757039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,2048,0.061549332406785756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,2048,0.036664889918433294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,2048,0.046315554115507335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,1536,0.047785778840382896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,1536,0.030124445756276447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,1024,0.03105866577890184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,1024,0.03449066811137729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,1024,0.023987554841571387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,1536,0.0377493335141076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,768,0.027914666467242773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,768,0.021500445074505273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,768,0.025936888323889837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,512,0.02055288851261139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,512,0.018915555543369718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,256,0.017273777061038546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,256,0.014341332846217685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,256,0.015915556086434256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,512,0.022095999783939783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,128,0.011058666639857821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,128,0.01407999959256914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,64,0.009297777381208207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,64,0.014157333307796054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,10240,32,0.009376000199053023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,10240,32,0.014793778459231058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,65536,0.7943769031100802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,10240,128,0.015440000428093804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,65536,1.5879750781589086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,65536,0.822623994615343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,16384,0.3738879892561171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,16384,0.20447556177775064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,16384,0.20270044273800322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,12288,0.2623644404941135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,12288,0.14975733227199978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,12288,0.15257955922020808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,10240,0.21873422463734946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,10240,0.12707377804650202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,10240,0.1296773354212443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,8192,0.17678399880727133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,8192,0.10334666570027669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,7168,0.09535911348130967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,7168,0.15593689017825657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,8192,0.10426755746205647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,7168,0.09099733167224461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,6144,0.13424800501929388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,6144,0.07860889037450154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,5120,0.07007110781139798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,5120,0.1128515534930759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,5120,0.06609421968460083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,6144,0.08001689116160075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,4096,0.09001333183712429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,4096,0.053152889013290405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,4096,0.05698933204015096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,3584,0.08039999670452543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,3584,0.047684444321526416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,3584,0.051073776351081006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,3072,0.06914844115575154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,3072,0.04155288802252875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,3072,0.0454053348965115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,2560,0.05908888578414917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,2560,0.036291556225882635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,2048,0.03388444582621256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,2048,0.04853244291411506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,2048,0.030598223209381104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,1536,0.02778488728735182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,1536,0.037736889388826154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,1536,0.02511111067401038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,2560,0.039523555172814265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,1024,0.026512889398468867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,1024,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,768,0.01982933282852173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,1024,0.022256000174416438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,768,0.021611554755104914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,768,0.018132444885041978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,512,0.016763556334707472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,512,0.01588000026014116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,256,0.01331555595000585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,256,0.01200977795653873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,256,0.013088888592190213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,256,0.008793777889675563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,128,0.00906577789121204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,128,0.011672889192899069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,64,0.007843555675612556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,64,0.0123848890264829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,8192,32,0.008040889269775814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,8192,32,0.012412444584899478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,512,0.016335111525323655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,65536,1.3991030587090387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,65536,0.8122835689120822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,8192,128,0.012423111332787408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,16384,0.2212675544950697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,16384,0.3266959985097249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,16384,0.20315378242068818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,12288,0.15244978004031712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,12288,0.2534862094455295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,12288,0.14976799488067627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,10240,0.20736977789137098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,10240,0.12568977144029406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,8192,0.10472710927327473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,65536,0.7917653189765083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,8192,0.1674168904622396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,8192,0.11290221744113499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,7168,0.16155021720462376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,7168,0.0960986680454678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,256,0.005501333210203383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,6144,0.12735733720991346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,10240,0.1290986670388116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,6144,0.07675821913613214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,5120,0.06917866733339098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,5120,0.107278221183353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,5120,0.06427288717693753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,4096,0.0870586633682251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,4096,0.052470223771201245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,3584,0.05191555619239807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,3584,0.07677066988415189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,3584,0.046839998828040234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,3072,0.045639998382992215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,7168,0.0923902193705241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,3072,0.06688533226648967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,3072,0.04110844598876105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,2560,0.05596977472305298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,2560,0.035642666949166193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,6144,0.0810684429274665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,2048,0.04595999916394552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,2048,0.030624889665179785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,256,0.006197333335876465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,1536,0.035418666071361966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,1536,0.0243848893377516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,4096,0.05820622046788534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,1024,0.026105778084860906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,1024,0.020139555136362713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,2560,0.0402728882100847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,768,0.021406221720907424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,768,0.018056000272432964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,1536,0.028372443384594385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,2048,0.03414222266938951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,512,0.016369778248998854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,512,0.015603555573357476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,256,0.011965333587593503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,256,0.013278221918476952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,1024,0.02362222141689724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,128,0.008719999757077958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,128,0.011391110718250275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,64,0.00756533361143536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,64,0.012054222325483957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,7168,32,0.007746666669845581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,7168,32,0.01218311074707243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,768,0.02035999960369534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,512,0.017696888910399545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,256,0.01354311075475481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,65536,1.1886719597710502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,16384,0.15724889437357584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,7168,128,0.012749333348539142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,16384,0.3227626747555203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,12288,0.11856622166103786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,16384,0.15446755621168348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,12288,0.1172506676779853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,12288,0.22552445199754503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,10240,0.17212977674272326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,10240,0.11132533020443386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,65536,0.6461911201477051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,8192,0.08136622111002605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,8192,0.14464978377024332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,8192,0.08972622288597955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,7168,0.07205600208706327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,7168,0.1243164406882392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,7168,0.07854311333762275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,6144,0.10681066910425822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,6144,0.06890133354398939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,5120,0.09341333309809367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,5120,0.05934222539265951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,65536,0.6063173082139757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,4096,0.07304355833265516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,4096,0.051572445366117686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,10240,0.10050666994518703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,3584,0.06307555569542779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,3584,0.03828622235192193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,5120,0.05487199955516391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,6144,0.06282489167319404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,3072,0.054986665646235146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,3072,0.03371644351217482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,2560,0.045984887414508395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,2560,0.02942755487230089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,2048,0.027285334136750963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,2048,0.03779200050565932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,2048,0.024816888901922438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,1536,0.022689777943823073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,1536,0.02920000089539422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,1536,0.020541333489947848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,4096,0.04494666722085741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,1024,0.020959999826219346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,1024,0.017318222257826064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,768,0.01684888866212633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,768,0.017204445269372728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,768,0.01556622154182858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,3584,0.04014844364590115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,512,0.013348444468445249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,512,0.013530666629473368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,3072,0.035615109735065036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,256,0.009600888523790572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,256,0.010968888799349466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,2560,0.03187288840611776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,128,0.007268444531493717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,128,0.010009778042634329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,64,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,64,0.0105608892109659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,6144,32,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,6144,32,0.01071999967098236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,1024,0.0188044442070855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,65536,0.9336729049682617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,512,0.014346667461925082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,65536,0.6410977575514052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,16384,0.22701509793599448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,16384,0.16327199671003553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,12288,0.10853422350353664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,12288,0.16625155342949763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,256,0.01218488895230823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,12288,0.11966221862369114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,10240,0.13911199569702148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,10240,0.10597155491511027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,8192,0.07643466525607638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,8192,0.11222310860951741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,8192,0.0815253323978848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,6144,128,0.010831111007266574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,7168,0.09870844417148167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,7168,0.0684808890024821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,65536,0.5532311333550347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,6144,0.09328266647126938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,6144,0.06966044505437215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,16384,0.13781600528293186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,5120,0.07225955857170953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,5120,0.05091377761628893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,10240,0.08816978004243638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,4096,0.05764000283347236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,4096,0.041759110159344144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,7168,0.06407911247677274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,3584,0.051207111941443555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,3584,0.03735111157099406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,3072,0.03218222326702542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,3072,0.04462044437726339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,3072,0.03322666552331712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,6144,0.055941330062018506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,2560,0.038005332152048744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,2560,0.02846222122510274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,5120,0.0484426650736067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,2048,0.031798223654429116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,2048,0.024525332782003615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,4096,0.04008177916208903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,1536,0.02462933295302921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,1536,0.02055644492308299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,3584,0.036347554789649114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,1024,0.018272888329293992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,1024,0.016957332690556843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,2560,0.028667555914984807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,768,0.015076445208655464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,768,0.015293333265516492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,2048,0.02481422159406874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,512,0.012040889097584618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,512,0.01258133351802826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,1536,0.020936000678274367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,256,0.008836444053384993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,256,0.010431999961535135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,1024,0.01696266730626424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,128,0.007072889142566257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,128,0.009761778016885122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,64,0.0061919999619325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,64,0.010341333018408881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,5120,32,0.006434666613737742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,5120,32,0.0105591110057301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,768,0.015243555108706156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,65536,0.7098399798075358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,512,0.013165333204799227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,256,0.011008888483047485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,65536,0.4845564630296495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,16384,0.176165329085456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,16384,0.1181173324584961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,5120,128,0.00977955593003167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,12288,0.1292195585038927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,12288,0.08402044243282741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,10240,0.1102293332417806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,10240,0.07152089145448473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,8192,0.0608160032166375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,8192,0.0890844464302063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,8192,0.05838844511244032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,65536,0.4369493325551351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,7168,0.08126666810777453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,7168,0.05429511268933614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,6144,0.049321777290768094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,6144,0.06787377595901489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,6144,0.04573511083920797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,5120,0.04100088940726386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,5120,0.05724622143639458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,5120,0.03932266765170627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,16384,0.11231911182403564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,4096,0.046075556013319224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,4096,0.031778666708204485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,12288,0.08569866418838501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,3584,0.0406986673672994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,3584,0.02847644355561998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,3072,0.02753244505988227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,3072,0.035402665535608925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,3072,0.02539288832081689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,10240,0.07276088661617704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,2560,0.030018667380015056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,2560,0.022451554735501606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,7168,0.05436799923578898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,2048,0.025306666890780132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,2048,0.019515555765893724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,4096,0.03384888834423489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,1536,0.020075556304719713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,1536,0.016770665844281513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,3584,0.03071111109521654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,1024,0.01497688889503479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,1024,0.013665777941544851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,2560,0.02499288817246755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,768,0.012602667013804117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,768,0.012460444536474017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,512,0.012171555724408893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,512,0.010119111173682744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,512,0.010619555910428366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,2048,0.021926222576035395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,256,0.00720177756415473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,256,0.008893333375453949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,1536,0.018584888842370775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,128,0.006103111224042044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,128,0.008431110945012834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,64,0.0053591111467944244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,64,0.008774222599135505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,4096,32,0.005504888792832692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,4096,32,0.008817777865462834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,1024,0.01533422205183241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,65536,0.6460408634609646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,768,0.01407199932469262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,65536,0.4620515505472819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,16384,0.16578843858506945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,16384,0.11192711194356282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,12288,0.08752355310651992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,12288,0.12424888875749375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,12288,0.08182311058044434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,256,0.010515555739402771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,10240,0.10322933064566718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,4096,128,0.009469333622190688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,10240,0.06927022006776598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,8192,0.05961600277158949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,8192,0.08360977967580159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,8192,0.056532442569732666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,7168,0.05350844396485222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,7168,0.07331466674804688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,7168,0.04972266819741991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,6144,0.0636062224706014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,6144,0.04337244563632541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,65536,0.4374524487389459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,5120,0.056714667214287647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,5120,0.04173333446184794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,16384,0.11229422357347275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,4096,0.04152889053026835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,4096,0.030996445152494643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,10240,0.07345155874888103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,3584,0.03694577680693732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,3584,0.02865333358446757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,6144,0.04677866564856636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,3072,0.03155555658870273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,3072,0.02498666610982683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,5120,0.04048888882001241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,2560,0.027464000715149775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,2560,0.022208889325459797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,4096,0.034767998589409724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,2048,0.02223555578125848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,2048,0.019392000304328073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,3584,0.03152799937460158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,1536,0.01761866609255473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,3072,0.028055111567179363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,1536,0.016498666670587327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,1024,0.01588800052801768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,1024,0.01294222225745519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,1024,0.013725333743625216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,768,0.014143110977278816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,768,0.010682666467295753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,768,0.012319111161761813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,512,0.008398222426573435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,512,0.009939555492666032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,2560,0.025337777203983728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,256,0.006191111273235745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,256,0.008924444516499838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,1536,0.01867111192809211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,128,0.0052986666560173035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,128,0.008392888638708327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,2048,0.021976000732845728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,64,0.004919111314747068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3584,32,0.005236444373925527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,64,0.00870577742656072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3584,32,0.008815999660227034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,65536,0.5605706638760036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,65536,0.34139643775092227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,16384,0.09015555514229669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,16384,0.1347057819366455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,16384,0.10884533325831096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,65536,0.4806266890631781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,12288,0.10410844617419773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,12288,0.08531200223498875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,512,0.012532444463835822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,10240,0.0863235526614719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,10240,0.06797155406739977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,8192,0.047932446002960205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,8192,0.06969955232408312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,8192,0.055626663896772593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,7168,0.04288088944223192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,7168,0.061425778600904674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,256,0.010576888918876648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,7168,0.04946577880117628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,6144,0.037610666619406805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,6144,0.052969776921802096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,6144,0.04304000073009067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,5120,0.03215466605292426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,5120,0.04437866806983948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,5120,0.036630223194758095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,4096,0.03642400105794271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,4096,0.030223111311594646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3584,128,0.009768000079525841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,3584,0.025082666012975905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,3584,0.03224444389343262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,3584,0.027627555860413447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,3072,0.02784444557295905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,3072,0.024898666474554274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,12288,0.06825511323081122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,2560,0.023815110325813293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,2560,0.022038223014937505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,10240,0.059225777784983315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,2048,0.02013777858681149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,2048,0.01938311093383365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,1536,0.015607999430762397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,1536,0.016232000456915963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,1536,0.01647733317481147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,4096,0.027759111589855615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,1024,0.012106666962305704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,1024,0.013624889155228933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,3072,0.023061333431137934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,768,0.010426667001512317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,768,0.011703111231327057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,2560,0.020610666937298246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,512,0.008421333299742805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,512,0.00962400022480223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,256,0.009542221824328104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,256,0.006124444305896759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,256,0.008760000268618265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,2048,0.018168000711335074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,1024,0.013349333571063148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,128,0.005156444592608346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,128,0.008441777692900764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,64,0.0046728890803125174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,64,0.008595555192894405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,3072,32,0.004896000027656555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,3072,32,0.008570666942331526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,65536,0.4534604284498427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,65536,0.443631993399726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,768,0.012112889024946423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,16384,0.11565155453152126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,16384,0.10607733329137166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,512,0.010945777926180096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,12288,0.08809510866800944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,12288,0.08054488897323608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,3072,128,0.008732444710201686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,10240,0.07295555538601346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,10240,0.0671022203233507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,65536,0.3020755449930827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,8192,0.05947022305594551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,8192,0.054791112740834556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,7168,0.03855289022127787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,7168,0.05243200063705444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,7168,0.0483457777235243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,16384,0.0810355544090271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,6144,0.045900444189707436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,6144,0.04302666584650675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,12288,0.06178310844633314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,5120,0.03869422276814779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,5120,0.03683999843067593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,10240,0.052355554368760854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,4096,0.03182933396763272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,4096,0.030606223477257624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,3584,0.023152889476882085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,3584,0.028267555766635474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,3584,0.0277164445983039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,8192,0.04297777679231432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,3072,0.02091288897726271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,3072,0.024522667129834492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,3072,0.024811555941899616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,2560,0.021213332811991375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,2560,0.021996445126003687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,6144,0.03390399946106805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,2048,0.01787911189926995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,2048,0.019119999474949308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,5120,0.029600001043743555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,1536,0.014913777510325113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,1536,0.01626755628320906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,4096,0.025182222326596577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,1024,0.011495999991893768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,1024,0.013331555657916598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,2560,0.018772444791264005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,2048,0.016692444682121277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,768,0.009949333137936061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,768,0.01124533348613315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,512,0.008084444536103142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,512,0.009641778137948778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,256,0.008797333472304875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,256,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,256,0.008805333740181392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,128,0.007825777762466008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,128,0.005026666654480828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,128,0.00816533300611708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,64,0.004554666578769684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,64,0.008551110823949179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2560,32,0.0047351109484831495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2560,32,0.008582221964995066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,65536,0.26737599902682835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,65536,0.3569191031985813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,65536,0.29302578502231175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,1536,0.014303111367755465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,16384,0.08979644378026326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,16384,0.06753422154320611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,12288,0.05458577805095249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,12288,0.06848355796602038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,12288,0.05152711272239685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,10240,0.046588444047504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,10240,0.057335111829969615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,10240,0.04407555527157254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,1024,0.01218577805492613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,8192,0.046893331739637584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,8192,0.037426667080985174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,768,0.011047999891969891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,7168,0.0410515566666921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,7168,0.032891554964913264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2560,512,0.010007999837398529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,6144,0.0352222224076589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,6144,0.02896889050801595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,16384,0.06984977589713202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,5120,0.0303439994653066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,5120,0.025824889540672302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,8192,0.03786399960517883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,4096,0.02460266649723053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,4096,0.021566222111384075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,7168,0.03440977798567878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,3584,0.02216266592343648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,3584,0.01995022263791826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,6144,0.029883556895785864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,3072,0.01924000018172794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,3072,0.018070222602950204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,5120,0.025935111774338618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,2560,0.01646133263905843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,2560,0.01644088824590047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,4096,0.0226248883538776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,2048,0.014103111293580798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,2048,0.014746665954589844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,1536,0.013087111214796701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,1536,0.011667555405033959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,1536,0.012974222501118978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,1024,0.011272889044549731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,1024,0.009168888959619734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,1024,0.010616888602574667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,3584,0.020804444948832195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,768,0.007902222375075022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,768,0.008993777963850234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,3072,0.01868977811601427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,512,0.0063893331421746155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,512,0.008079110748238033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,2560,0.01700355609258016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,256,0.004940444396601783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,256,0.007370666497283512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,2048,0.015065777632925244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,128,0.00435288871328036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,128,0.007043555378913879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,64,0.003975999852021535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,768,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,64,0.007346666521496243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,2048,32,0.0041537777417235905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,2048,32,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,65536,0.25777334637112087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,65536,0.2730097770690918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,65536,0.26812977261013454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,16384,0.07089244657092623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,16384,0.06428888771269056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,12288,0.05306755503018697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,12288,0.05355199840333727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,12288,0.04954666561550564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,10240,0.04534577661090427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,10240,0.0447635551293691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,10240,0.04283644424544441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,512,0.008997333546479544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,8192,0.036002665758132935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,8192,0.03571110963821411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,7168,0.03378666771782769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,7168,0.031825777557161115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,7168,0.03194133440653483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,256,0.007536888950400882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,6144,0.02792088852988349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,6144,0.028508444627126057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,2048,128,0.006761777732107375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,5120,0.023623999622133043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,5120,0.025261332591374714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,16384,0.0686088866657681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,4096,0.01921066641807556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,4096,0.021292444732454088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,6144,0.029370665550231934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,3584,0.017322666115230985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,3584,0.019553777244355943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,8192,0.03747377792994181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,3072,0.015372445185979208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,3072,0.017830222845077515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,2560,0.016794666647911072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,2560,0.013456888496875763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,5120,0.02586666742960612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,2560,0.016207999653286405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,2048,0.011504000259770287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,2048,0.014484445254007975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,4096,0.022112889422310725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,1536,0.009653333160612319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,1536,0.012718222207493253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,3584,0.020421332783169217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,1024,0.007595555649863348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,1024,0.00981244444847107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,768,0.009577777650621202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,768,0.006321777900060018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,768,0.008854221966531541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,3072,0.018251554833518136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,512,0.005366222312053044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,512,0.007853333320882585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,256,0.006783999916579988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,256,0.004480000171396467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,256,0.007265778051482306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,128,0.006203555398517185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,128,0.003932444585694207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,128,0.006894222150246303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,2048,0.014783110883500842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,64,0.0036151111125946045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1536,32,0.0037662221325768363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,64,0.007075555622577667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1536,32,0.007111999723646376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,1536,0.01257333325015174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,65536,0.18533510631985137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,16384,0.05522222320238749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,16384,0.052320887645085655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,65536,0.25782132148742676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,16384,0.0639573335647583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,12288,0.04456622070736355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,12288,0.04401422209209866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,12288,0.04942933387226529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,10240,0.04156088829040527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,10240,0.04244355691803826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,1024,0.010984000232484607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,8192,0.03122044603029887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,8192,0.03484888871510824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1536,512,0.00810222244924969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,7168,0.026791110634803772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,7168,0.03113244308365716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,65536,0.20873955885569254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,6144,0.023552889625231426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,6144,0.027694221999910142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,10240,0.03867910967932807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,5120,0.02261688808600108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,5120,0.024477332830429077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,4096,0.018983999888102215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,4096,0.015944888194402058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,4096,0.020675554871559143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,8192,0.03062844607565138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,3584,0.014635556273990206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,3584,0.01901155544651879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,3072,0.015654222832785714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,3072,0.01385600037044949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,3072,0.01735555628935496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,7168,0.027841776609420776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,2560,0.011864888999197217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,2560,0.015754666593339708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,6144,0.0251884443892373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,2048,0.009935110807418823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,2048,0.013935110635227628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,5120,0.021714667479197185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,1536,0.008404444489214156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,1536,0.012348444097571902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,3584,0.017491555876202054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,1024,0.006112888869312074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,1024,0.009413333402739631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,768,0.007480888730949826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,768,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,768,0.00868711123863856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,512,0.006792888873153263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,512,0.004607999904288186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,512,0.007712889048788283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,2560,0.014570666684044732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,256,0.003950222084919612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,256,0.007120889094140794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,128,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,128,0.003567999849716822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,128,0.006754666566848755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,64,0.0032879999942249725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,64,0.006893333461549547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,1024,32,0.0033857776886887024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,1024,32,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,2048,0.012841777669058906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,65536,0.14435199896494547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,65536,0.25816443231370717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,1536,0.01076533314254549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,16384,0.041344000233544245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,16384,0.06414755847718981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,1024,0.008416000339719985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,12288,0.03525866733656989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,12288,0.049230221245023936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,1024,256,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,10240,0.03528177738189697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,10240,0.04176266656981575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,8192,0.030287110143237646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,8192,0.026693332526418898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,8192,0.03436444534195794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,65536,0.21215911706288657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,7168,0.02439111140039232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,7168,0.03128622306717767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,6144,0.024001777172088623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,6144,0.02088977727625105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,6144,0.02776088813940684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,16384,0.05800977680418226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,5120,0.01714933415253957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,5120,0.024757333927684363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,4096,0.018195556269751657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,4096,0.014527999692493014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,4096,0.020759110649426777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,12288,0.0443466669983334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,3584,0.013415111435784234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,3584,0.01905955539809333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,10240,0.03716622127426995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,3072,0.012343111137549082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,3072,0.01724088854259915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,7168,0.027496889233589172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,2560,0.010287111004193624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,2560,0.01568888955646091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,2048,0.012253333297040729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,2048,0.00908622228437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,2048,0.013964444398880005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,5120,0.02120888895458645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,1536,0.007621333003044128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,1536,0.01199288914601008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,1024,0.007802666889296637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,1024,0.005384888913896348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,1024,0.009336888790130615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,3584,0.016662221815851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,768,0.0047280001971456735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,768,0.008622222476535374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,512,0.0064479998416370815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,512,0.004092444562249714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,512,0.007791111038790808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,3072,0.01512888901763492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,256,0.0036551112102137674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,256,0.007241778075695038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,2560,0.013649777405791812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,128,0.003256888853179084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,128,0.006786666810512543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,64,0.0030622221529483795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,64,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,768,32,0.0032088889016045462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,768,32,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,1536,0.00979555563794242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,65536,0.1003555523024665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,65536,0.2554088963402642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,768,0.007254222200976477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,16384,0.03267822331852383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,16384,0.0643120010693868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,256,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,12288,0.029133333100212946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,12288,0.04890311095449659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,768,128,0.005519111123349931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,10240,0.029822223716311987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,10240,0.04137333234151205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,65536,0.17593867248959014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,8192,0.01906044450071123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,16384,0.047835555341508657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,8192,0.03433777888615926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,7168,0.020263110597928364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,7168,0.03099200129508972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,12288,0.037434667348861694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,6144,0.017818666166729398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,6144,0.02753244505988227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,5120,0.01850044396188524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,5120,0.015656888484954834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,5120,0.02455555564827389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,4096,0.016015999846988253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,4096,0.011039999624093374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,4096,0.02067377832200792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,3584,0.01481244464715322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,3584,0.010877333581447601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,3584,0.018858666221300762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,10240,0.03223022156291538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,3072,0.010383110907342698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,3072,0.01718933383623759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,8192,0.026224000586403742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,2560,0.008790222307046255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,2560,0.015591111448076038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,7168,0.023649776975313824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,2048,0.007575111256705389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,2048,0.013919111755159167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,1536,0.008593777815500895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,1536,0.005693333430422678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,1536,0.011175110936164856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,6144,0.02128533356719547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,1024,0.0047608887155850725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,1024,0.009054222040706212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,3072,0.013424889081054263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,768,0.004311999926964442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,768,0.008460444708665213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,512,0.006033777776691649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,512,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,512,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,2560,0.012355555262830524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,256,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,256,0.003412444351447953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,256,0.007083555890454187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,128,0.0031537777847713898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,128,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,64,0.0030008889734745026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,64,0.00683111117945777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,512,32,0.003128888913326793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,512,32,0.00685066646999783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,65536,0.14785510963863796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,65536,0.0667111078898112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,65536,0.25677333937750924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,2048,0.010749333434634738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,16384,0.022915555371178523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,16384,0.06457600328657362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,1024,0.007159111400445302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,12288,0.020392888122134738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,12288,0.04888088835610283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,10240,0.027128888501061335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,10240,0.018996444013383653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,10240,0.041391998529434204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,8192,0.022652443912294176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,8192,0.013833777772055732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,768,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,8192,0.034224887688954674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,7168,0.013983110586802164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,7168,0.031207998593648274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,512,128,0.005154666801293691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,6144,0.014724445011880664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,6144,0.02756799923049079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,16384,0.04124177826775445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,5120,0.013022222452693515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,5120,0.024342222346199885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,4096,0.014255111416180929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,4096,0.010807111031479306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,4096,0.020581333173645865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,3584,0.013240888714790344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,3584,0.009730666875839233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,3584,0.0188719994492001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,12288,0.03192977772818671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,3072,0.008973333570692275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,3072,0.017151999804708693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,7168,0.020473778247833252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,6144,0.01835644410716163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,2560,0.007681777907742395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,2560,0.0155404441886478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,2048,0.00684799998998642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,2048,0.013691555294725629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,5120,0.01627377834584978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,1536,0.005761777775155173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,1536,0.011160000330872007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,3072,0.012053333222866058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,1024,0.0047866664826869965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,1024,0.008945778012275696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,2560,0.010562666588359408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,768,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,768,0.004313777718279096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,768,0.008426666259765625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,512,0.0037724446091387006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,512,0.0058124446206622655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,512,0.007600888609886169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,256,0.00327822214199437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,256,0.00703999979628457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,128,0.005062222066852781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,128,0.003056888985964987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,128,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,64,0.0029164445069101122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,64,0.006746666712893381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,256,32,0.0029306666304667792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,256,32,0.00674044465025266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,65536,0.14622578356001112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,65536,0.05652799871232775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,1536,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,65536,0.25653600692749023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,2048,0.009334222310119206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,16384,0.019156444403860304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,16384,0.06406222449408637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,12288,0.0155395550860299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,12288,0.04811822374661764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,1024,0.006758222149478064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,10240,0.013775111072593264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,10240,0.041377776198916964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,256,256,0.005467555589146084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,8192,0.012075555821259817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,8192,0.03484266665246751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,7168,0.020188444190555148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,7168,0.01162755572133594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,7168,0.031473779016070895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,16384,0.04034666551484002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,6144,0.013378666506873237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,6144,0.027592889136738245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,12288,0.030986666679382324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,5120,0.009927999642160203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,5120,0.024417777856190998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,4096,0.013944889108339945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,4096,0.009474666582213508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,4096,0.020703110429975722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,10240,0.02679288884003957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,3584,0.008199111455016667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,3584,0.019102222389645047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,8192,0.022164444128672283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,3072,0.007652444144090016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,3072,0.017238222890430026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,6144,0.0181040002240075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,2560,0.006834666762087081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,2560,0.015676443775494892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,5120,0.016047999262809753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,2048,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,2048,0.013222222526868185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,3584,0.012894222305880653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,1536,0.00490488898422983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,1536,0.010517333116796283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,1024,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,1024,0.0041884444653987885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,1024,0.008930666579140557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,768,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,768,0.0038808890514903595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,768,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,3072,0.01166311071978675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,512,0.0035333335399627686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,512,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,2560,0.010472888747851053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,256,0.003152888889114062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,256,0.0070426662762959795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,2048,0.009097778134875828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,128,0.002941333378354708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,128,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,64,0.0028435554769304064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,64,0.006765333314736684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,128,32,0.0029164445069101122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,128,32,0.006771555377377405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,65536,0.05622400177849663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,65536,0.2561555438571506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,16384,0.01796266602145301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,16384,0.0644577807850308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,12288,0.014579556054539151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,12288,0.04829066660669115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,10240,0.013191999660597907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,10240,0.04124710957209269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,8192,0.011295999917719098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,8192,0.03455200129085117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,7168,0.010653333531485664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,7168,0.0313297775056627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,6144,0.010675555302037133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,6144,0.027415111660957336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,5120,0.009684444301658207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,5120,0.02424888809521993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,4096,0.008960000342792934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,4096,0.02055288851261139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,3584,0.008120888637171851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,3584,0.01896711190541585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,3072,0.007542221910423703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,3072,0.017119111286269296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,2560,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,2560,0.015396444333924187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,2048,0.005708444449636672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,2048,0.013112000293201871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,1536,0.0047644442982143825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,1536,0.010443555812040964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,1024,0.004109333372778363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,1024,0.008887111312813228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,768,0.0037804444630940757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,768,0.008406221866607666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,512,0.0034604445099830627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,512,0.007613333563009898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,256,0.003117333269781537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,1536,0.007805333369308048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,256,0.007085333267847697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,128,0.0028577778074476454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,128,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,64,0.006671111202902264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,64,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,64,32,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,64,32,0.0029333333174387612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,65536,0.05384000142415365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,65536,0.25549599859449595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,16384,0.01705777810679542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,12288,0.014193778236707052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,16384,0.06412088871002197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,12288,0.04820355441835192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,10240,0.012896888785892062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,8192,0.011181332998805575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,10240,0.0413973331451416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,7168,0.010444444086816577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,8192,0.03478666808870103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,512,0.005772444523043103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,6144,0.009542221824328104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,7168,0.031425777408811785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,5120,0.009204444785912832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,5120,0.02424000038041009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,6144,0.027450667487250432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,4096,0.008455110920800103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,3584,0.008216000265545314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,4096,0.020568889048364427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,3584,0.01883555617597368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,3072,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,3072,0.017070222232076857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,2560,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,2560,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,2048,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,2048,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,1536,0.004668444395065308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,1536,0.010398222340477837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,1024,0.004032888760169347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,1024,0.008940444224410588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,768,0.0036640001667870414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,512,0.003335111050142182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,768,0.008406221866607666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,512,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,256,0.002979555477698644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,256,0.007056000332037608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,128,0.0027679999669392905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,128,0.006554666492674086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,64,0.0026853332916895547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,64,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1024,32,32,0.002620444529586368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1024,32,32,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,256,0.005432888865470886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,16384,1.1351946724785698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,16384,1.1931439505683052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,16384,2.5460327996148004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,12288,1.8990160624186199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,12288,0.9270915985107422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,10240,0.777523570590549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,10240,1.5659991370307074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,8192,1.2251013649834526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,8192,0.6106293466356065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,12288,0.8439893192715116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,7168,0.5082595613267686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,7168,0.9762836032443576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,10240,0.7251066631740994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,7168,0.5433653195699056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,6144,0.9411715401543511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,6144,0.4571777979532878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,5120,0.7894701957702637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,5120,0.38234400749206543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,8192,0.5856684578789605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,4096,0.30103733804490834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,4096,0.6367448700798882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,4096,0.3153457906511095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,3584,0.525526205698649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,3584,0.26771643426683217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,3072,0.22969243261549208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,3072,0.4347662130991618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,6144,0.43403644031948513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,3072,0.2309786743587918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,2560,0.21847111648983428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,2560,0.3700106673770481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,2560,0.20015022489759657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,2048,0.16908444298638237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,2048,0.29004356596204967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,5120,0.37442755699157715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,2048,0.1591671175426907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,1536,0.1282880041334364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,1536,0.2513048913743761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,1536,0.12073866526285808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,1024,0.17703289455837676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,1024,0.09494755665461223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,768,0.1229413350423177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,768,0.08312177658081055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,512,0.06063999732335409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,512,0.088319996992747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,3584,0.26215823491414386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,512,0.07748355468114217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,256,0.05834488736258613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,256,0.06033511294258965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,128,0.04910666743914286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,128,0.050435556305779346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,128,0.05713599920272827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,64,0.04500444398985969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,64,0.05821688969930013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,65536,32,0.04552533229192098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,65536,32,0.05843289030922783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,1024,0.09199821949005127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,768,0.0772213339805603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,65536,1.2264027065700955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,16384,0.2974969016181098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,16384,0.5597644382052952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,16384,0.30180178748236763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,65536,2.4464213053385415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,65536,1.259176042344835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,65536,256,0.055697778860727944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,12288,0.4131440056694879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,12288,0.23633866839938691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,10240,0.1855377753575643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,10240,0.3641013304392497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,8192,0.28424978256225586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,8192,0.1542044480641683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,8192,0.1671137809753418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,7168,0.2485048770904541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,7168,0.1358328925238715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,12288,0.22334843211703828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,6144,0.20541955365075004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,6144,0.11490221818288167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,5120,0.10115110874176025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,10240,0.1852302286359999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,5120,0.1733671161863539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,5120,0.09522222148047553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,4096,0.13698400391472712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,4096,0.0783546699417962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,7168,0.13649245103200278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,3584,0.12343645095825195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,3584,0.06951022148132324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,6144,0.1162702242533366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,3072,0.10628444618648952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,3072,0.06085511048634847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,2560,0.054254220591651074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,2560,0.08995288610458374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,2560,0.05268889003329807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,4096,0.0816826687918769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,2048,0.07426933447519939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,2048,0.04377866784731547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,1536,0.03746044304635789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,1536,0.05790311098098755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,1536,0.03558933403756883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,3584,0.07218399975034925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,1024,0.02913244565327962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,1024,0.0410791105694241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,1024,0.02863111098607381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,3072,0.06363022327423096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,768,0.03380266825358073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,768,0.02536711096763611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,512,0.02103999919361538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,512,0.026201777988009985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,512,0.022452443838119507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,256,0.01660355594423082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,256,0.018735999862353008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,256,0.01866755551762051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,128,0.014448889427714877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,128,0.017383111847771537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,64,0.0127697777416971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,64,0.016523554921150208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,16384,32,0.013176889055305056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,16384,32,0.0169813334941864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,2048,0.04546844296985202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,65536,0.9134222666422526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,65536,1.7773449156019423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,16384,0.22848267025417754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,65536,0.9911777708265516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,16384,0.4032488928900824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,16384,0.25036001205444336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,12288,0.3015057775709364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,12288,0.1923431158065796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,10240,0.14691555500030518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,10240,0.25068532096015084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,768,0.025369778275489807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,16384,128,0.015180443723996481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,10240,0.16159910625881618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,8192,0.2006373273001777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,8192,0.12477510505252415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,7168,0.1035368906127082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,7168,0.17645155058966744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,7168,0.10631199677785237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,6144,0.09017244312498306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,6144,0.15247644318474665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,6144,0.09709688689973618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,5120,0.12786044014824763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,5120,0.08135822084214953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,12288,0.16814222600724962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,4096,0.10200977987713283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,4096,0.06529600090450711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,3584,0.055368890364964805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,8192,0.11390133698781331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,3584,0.08945600191752116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,3584,0.05677422549989489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,5120,0.07629866732491387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,3072,0.07892800039715238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,3072,0.05177066723505656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,4096,0.06203999784257677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,2560,0.06658399767345853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,2560,0.04425866736306084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,2048,0.053872889942593045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,2048,0.03713688916630215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,1536,0.029834667841593426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,1536,0.04225688841607836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,1536,0.030336889955732558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,1024,0.02369244396686554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,1024,0.030223998758527968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,1024,0.023779556155204773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,768,0.02128622266981337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,768,0.024238222175174292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,768,0.02106933295726776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,512,0.017849778135617573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,512,0.018739556272824604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,512,0.018597332967652213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,256,0.01458755632241567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,256,0.01331022216214074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,256,0.015503999259736804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,3072,0.04832444588343302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,128,0.010205333431561789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,128,0.01332266628742218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,64,0.008353778057628209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,64,0.013786666923099093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,12288,32,0.008652444514963362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,12288,32,0.01442311041884952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,2560,0.042845331960254245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,2048,0.035679999325010515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,65536,1.5236480500962999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,12288,128,0.012961777547995249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,16384,0.2146986722946167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,65536,0.8096302350362142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,16384,0.33673776520623105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,16384,0.1985937754313151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,12288,0.15022044711642796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,12288,0.2747173309326172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,12288,0.15524621804555258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,10240,0.12711911731296116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,10240,0.2306577894422743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,8192,0.10292177730136448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,10240,0.12995466921064588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,8192,0.1898124482896593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,8192,0.09965333011415269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,7168,0.1653439998626709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,7168,0.08824711375766331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,6144,0.08521333005693223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,6144,0.13145244121551514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,6144,0.0819973349571228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,5120,0.06776533524195354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,5120,0.11292800638410781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,5120,0.06638044118881226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,4096,0.08945333295398289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,4096,0.053598221805360585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,3584,0.050274666812684804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,3584,0.07912621895472209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,3584,0.04782399866316053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,3072,0.04430844386418661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,3072,0.06853689087761773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,3072,0.04270044300291273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,65536,0.7851200103759766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,2560,0.06624266836378309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,2560,0.04215111003981697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,2048,0.03784266776508755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,2048,0.05445333321889242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,2048,0.03567911187807719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,7168,0.09198577536476983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,1536,0.037682665718926325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,1536,0.025343111819691126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,4096,0.05531644490030077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,1024,0.026720888084835474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,1024,0.020337777005301583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,768,0.01973600023322635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,768,0.021800888909233943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,768,0.018256889449225534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,2560,0.03872888949182298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,512,0.016756445169448853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,512,0.01591377788119846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,256,0.013384888569513956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,256,0.011879111329714457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,256,0.01332711097266939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,1536,0.02736888825893402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,128,0.012531555361217923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,128,0.011928000383906893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,64,0.009778666827413771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,64,0.012429333395428128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,10240,32,0.01037333326207267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,10240,32,0.012482666307025485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,1024,0.02190133266978794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,65536,1.1997253629896376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,512,0.016350222958458793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,16384,0.1537448830074734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,16384,0.2718844413757324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,10240,128,0.012050666742854647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,16384,0.1530204481548733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,65536,0.6278364393446181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,12288,0.11734310785929362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,12288,0.2212186654408773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,12288,0.11464088492923313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,10240,0.1716853380203247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8192,128,128,0.0064195555945237475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,10240,0.09850133127636379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,8192,0.1373466650644938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,8192,0.08127022451824613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,7168,0.07349777883953519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,7168,0.12124621868133545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,7168,0.0702435572942098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,6144,0.06250400013393827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,6144,0.1043155524465773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,6144,0.060673779911465116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,5120,0.0532968905236986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,5120,0.08756888575024074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,5120,0.051346666283077665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,4096,0.045421332120895386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,4096,0.07123555739720662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,4096,0.04330844349331326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,3584,0.03996977872318692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,3584,0.06300266583760579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,3584,0.038557334078682795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,3072,0.03566311134232415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,3072,0.05478399991989136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,3072,0.03447022371821933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,2560,0.031189332405726116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,2560,0.04636711213323805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,2560,0.03011022342575921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,2048,0.03790311018625895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,2048,0.025432889660199482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,1536,0.0229120006163915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,1536,0.029495110114415485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,1536,0.020855110552575853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,1024,0.018481777773963082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,1024,0.021207999851968553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,1024,0.017143110434214275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,768,0.016798221402698092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,768,0.01737600068251292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,768,0.015433778365453085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,512,0.014333332578341166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,512,0.01352266636159685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,512,0.013636444177892474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,256,0.012038222617573209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,256,0.009764444496896531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,256,0.011117333339320289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,65536,0.605604436662462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,128,0.008751110898123847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,128,0.011726222104496427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,64,0.007479999628331926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,64,0.012200888660218982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,8192,32,0.007744000189834171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,8192,32,0.01256533298227522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,10240,0.10069421927134196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,8192,0.08238666587405734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,65536,1.0039493772718642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,65536,0.6213733355204264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,16384,0.24731021457248262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,16384,0.1552542183134291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4096,128,128,0.005523555394675996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,12288,0.1796435515085856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,2048,0.02657866643534766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,12288,0.11339555846320258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,10240,0.15080889066060385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,10240,0.09805599848429362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,8192,128,0.010871111518806882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,8192,0.1213466723759969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,8192,0.07689422369003296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,7168,0.0683493349287245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,7168,0.10701155662536621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,7168,0.06763111220465766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,65536,0.5751084221733941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,6144,0.09679555230670506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,6144,0.06586755646599664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,16384,0.14945067299736872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,5120,0.07805599768956502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,5120,0.05161866545677185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,4096,0.04322133461634318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,4096,0.06337422132492065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,4096,0.04222044348716736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2048,128,128,0.005249777601824866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,12288,0.11534044477674697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,3584,0.05540088812510172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,3584,0.03849422269397312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,3072,0.04950933323966133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,3072,0.03345155715942383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,10240,0.09386933512157863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,2560,0.04026222229003906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,2560,0.029170665476057265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,2048,0.02587644424703386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,2048,0.03291644321547614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,2048,0.024461333950360615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,1536,0.02236088944805993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,8192,0.07694844404856364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,1536,0.025769778423839148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,1536,0.020417778028382193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,1024,0.018168000711335074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,1024,0.018522666560279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,1024,0.017044444878896076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,768,0.016318221886952717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,768,0.015192889504962497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,768,0.015127999915017022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,6144,0.059393776787651904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,512,0.011678222152921887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,512,0.013386666774749756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,256,0.011896889242861005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,256,0.008989333278603023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,256,0.010590222146775989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,128,0.010473777850468954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,128,0.007847111258241866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,128,0.00978488889005449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,64,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,64,0.010266666611035665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,7168,32,0.006478222293986215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,7168,32,0.010478222535716163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,5120,0.050761779149373375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,65536,0.8158933321634928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,65536,0.5879831314086914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,16384,0.19600533114539254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,16384,0.15087821748521593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,3584,0.038839999172422625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,12288,0.1488071150249905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,12288,0.11201243930392796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,3072,0.03420711225933499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,2560,0.030329777134789362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,10240,0.12444978290134007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,10240,0.09347822268803914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,8192,0.0783768892288208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,8192,0.10272355874379475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,7168,0.07146044572194417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,8192,0.07539555761549208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,7168,0.09121421972910564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,7168,0.0669671098391215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,6144,0.07645866605970594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,6144,0.05912444326612684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,5120,0.05204000075658163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,5120,0.0646631121635437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,7168,512,0.014095111025704278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,5120,0.05028088887532552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,4096,0.052594665024015635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,4096,0.041109333435694374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1024,128,128,0.005126222140259213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,3584,0.045992887682384916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,3584,0.037055111593670316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,3072,0.03419466813405355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,3072,0.040116445885764226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,3072,0.032872888776991106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,65536,0.6069768799675835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,2560,0.03518844313091702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,2560,0.0299635562631819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,16384,0.1500062280231052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,2048,0.02812533246146308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,2048,0.02420088814364539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,12288,0.11437955167558457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,1536,0.022612444228596155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,1536,0.020319110817379422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,1024,0.018583999739752877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,1024,0.016707556115256417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,1024,0.017045333981513977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,10240,0.09702577855851914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,768,0.014093332820468478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,768,0.015227556228637695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,512,0.014002665877342224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,512,0.01126577787929111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,512,0.013018666870064206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,6144,0.06174577607048882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,256,0.008135110967689091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,256,0.010469333165221743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,4096,0.04282044370969137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,128,0.006517333288987477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,128,0.009630222287442949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,64,0.005690666536490123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,64,0.010095111197895473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,6144,32,0.005975111077229182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,6144,32,0.01038755559258991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,3584,0.03937777876853943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,65536,0.7052888870239258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,2560,0.030551999807357788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,65536,0.44094310866461855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,16384,0.1846942239337497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,2048,0.026370666093296472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,16384,0.11661156018575032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,12288,0.08201866679721408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,12288,0.14336266782548693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,10240,0.07636088795132108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,12288,0.08045688602659437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,10240,0.10918933153152466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,10240,0.06865600082609388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,8192,0.0874648888905843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,8192,0.05910578038957384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,8192,0.05484177668889364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,7168,0.07697333229912652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,7168,0.049138665199279785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,1536,0.022527111901177302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,6144,0.06572444571389092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,6144,0.043022221989101835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,768,0.016528000434239704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,5120,0.05592978000640869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,256,0.011626666618718041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,5120,0.03765689002143012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,6144,128,0.010232888989978367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,4096,0.04658133453792996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,4096,0.030918220678965252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,3584,0.040349334478378296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,3584,0.02827199962404039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,65536,0.4067591031392415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,3072,0.03585422370168898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,3072,0.025059555967648823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,16384,0.10669155253304376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,2560,0.024469332562552557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,2560,0.022293332550260756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,2560,0.030016889174779255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,2048,0.024863999750879075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,2048,0.02130577796035343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,2048,0.018984888990720112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,1536,0.01994400057527754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,1536,0.016191111670600045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,7168,0.051381333006752856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,1024,0.015339555011855232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,1024,0.013835555149449242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,6144,0.0446337792608473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,768,0.012850667039553324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,768,0.012511110968059964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,5120,0.038400888442993164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,512,0.010263111028406356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,512,0.010144000252087912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,4096,0.032905777295430504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,256,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,256,0.008807111117574904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,128,0.009464888936943477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,128,0.006155555446942647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,128,0.008237332933478886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,64,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,64,0.008744888835483128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,5120,32,0.005400889035728242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,5120,32,0.008805333740181392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,3584,0.02990666694111294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,3072,0.02696711156103346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,65536,0.5717422167460123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,65536,0.42814045482211643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,16384,0.13592533270517984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,16384,0.10258755418989395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,12288,0.0657528903749254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,12288,0.10299644205305312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,12288,0.07801066504584418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,1024,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,1536,0.0181013329161538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,10240,0.0557617810037401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,10240,0.08617155419455634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,10240,0.0653653343518575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,8192,0.045808000697029956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,8192,0.0695742236243354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,8192,0.0544088880221049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,7168,0.0611635579003228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,7168,0.04796977837880453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,6144,0.052579555246565074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,6144,0.04204533497492472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,768,0.013934221532609729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,512,0.012306667036480375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,5120,0.0449022220240699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,5120,0.03724177678426107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,4096,0.026910222238964502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,4096,0.03638133406639099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,4096,0.030294222964180842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,3584,0.03234666585922241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,3584,0.02481777800454034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,3584,0.027992000182469685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,3072,0.022595556245909795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,3072,0.028251555230882432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,3072,0.024784000383483037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,2560,0.02036177780893114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,2560,0.024015110399987962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,2560,0.02188266648186578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,2048,0.017477333545684814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,2048,0.020060444871584576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,2048,0.01907911068863339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,1536,0.015296889675988091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,1536,0.016072000066439312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,1536,0.016167999969588388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,1024,0.013046222428480784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,1024,0.012270222107569376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,1024,0.013603555659453074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,768,0.012170666621790992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,768,0.010404444403118556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,768,0.01181244436237547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,512,0.010954666468832227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,5120,256,0.010461333725187512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,512,0.008219555848174626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,512,0.00963466697269016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,256,0.006115555349323485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,256,0.008658666577604082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,128,0.00869511150651508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,128,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,128,0.008071111308203803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,64,0.004593777573770947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,64,0.008524444368150499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,4096,32,0.004847110973464118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,4096,32,0.008638222184446123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,65536,0.3142862319946289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,65536,0.46857690811157227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,65536,0.33391110102335614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,16384,0.08874755435519749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,16384,0.12019377284579807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,65536,0.4300862153371175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,16384,0.11183555920918782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,12288,0.09127377801471287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,12288,0.0645680030186971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,12288,0.08462755547629462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,10240,0.05483466717931959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,10240,0.07675644424226549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,10240,0.06647555695639716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,8192,0.06143199735217624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,8192,0.0543813341193729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,7168,0.05417066812515259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,16384,0.08811466561423408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,7168,0.04828088813357883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,6144,0.04687111245261299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,6144,0.04158400164710151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,5120,0.030794666873084173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,5120,0.03976088762283325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,5120,0.03627999954753452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,7168,0.040947556495666504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,4096,0.03233955634964837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,4096,0.02985066506597731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,6144,0.03596355517705282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,3584,0.028953777419196233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,3584,0.027264888087908428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,5120,0.03122488988770379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,3072,0.02514666650030348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,3072,0.02442666722668542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,2560,0.01958666741847992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,2560,0.021536888347731695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,2560,0.02165866725974613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,2048,0.01702933344576094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,2048,0.018032888571421306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,2048,0.018710222509172227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,1536,0.015118221441904703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,1536,0.014950222439236112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,1536,0.01606044504377577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,4096,256,0.00962311112218433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,1024,0.011711999773979187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,1024,0.01328888866636488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,8192,0.04503555430306328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,768,0.010272000398900773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,768,0.01127911110719045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,6144,0.03519377774662442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,7168,0.03993689020474752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,512,0.008347555167145198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,512,0.00962488849957784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,256,0.006451555424266391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,256,0.00870666652917862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,128,0.008392000363932716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,128,0.005691555639108022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,128,0.008059555457697975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,64,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,64,0.008379555410808986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3584,32,0.005390222287840313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3584,32,0.008595555192894405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,4096,0.02623733381430308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,65536,0.4024026658799913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,65536,0.3148533238304986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,3584,0.024138665861553613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,3072,0.021998221675554912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,16384,0.10316089126798843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,65536,0.4205288887023926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,12288,0.062296887238820396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,16384,0.10085244311226739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,12288,0.07843110958735149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,12288,0.07657866345511542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,10240,0.06573777728610568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,8192,0.04459111226929558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,10240,0.0646524429321289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,8192,0.053020444181230336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,8192,0.05250933435228136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,7168,0.04010399844911363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,7168,0.046741333272722035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,6144,0.03479911221398248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,7168,0.04790222313668993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,6144,0.040384888648986816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,6144,0.04175466630193922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,5120,0.034323556555642024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,5120,0.03626755542225308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,1024,0.012793777717484368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,4096,0.027661333481470745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,4096,0.029767112599478826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,3584,0.024088000257809956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,3584,0.024714666936132643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,3584,0.027165333429972332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,768,0.011914666328165265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,3072,0.021499555971887376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,3072,0.02429155508677165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,2560,0.01978577839003669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,2560,0.018446221947669983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,2560,0.021631111701329548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,512,0.010727110836240979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,2048,0.015792000624868605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,2048,0.018735999862353008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,1536,0.01498133275243971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3584,256,0.009352888498041365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,1536,0.012816000315878125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,1536,0.015863999724388123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,16384,0.08277777830759685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,1024,0.010010666317409938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,1024,0.013031110995345645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,768,0.011904888682895236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,768,0.008637333081828224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,768,0.010763555765151978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,512,0.010688888529936472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,512,0.007177777588367462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,512,0.009482666850090027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,10240,0.053299556175867714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,256,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,256,0.008520888785521189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,128,0.00461422238085005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,128,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,64,0.0041697778635554844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,64,0.008264000217119852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,3072,32,0.004354666504595014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,3072,32,0.00830577810605367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,5120,0.030495110485288832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,4096,0.026234666506449383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,65536,0.34888090027703184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,3072,0.02176888949341244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,65536,0.2533431053161621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,16384,0.08800711234410603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,16384,0.06129778093761868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,12288,0.06781955560048421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,12288,0.04717600014474657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,10240,0.042279998461405434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,10240,0.056868447197808154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,10240,0.04053777787420485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,2048,0.01703466640578376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,8192,0.04539022180769178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,8192,0.03377066718207465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,7168,0.03154755632082621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,7168,0.04008355405595567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,7168,0.030626667870415583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,1024,0.012906666431162091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,6144,0.03610400027698941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,6144,0.0273644444015291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,256,0.008987555901209513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,5120,0.029872000217437744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,3072,128,0.008070222205585903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,5120,0.02404444416364034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,4096,0.02455644475089179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,4096,0.02045866681469811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,65536,0.24612800280253092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,3584,0.021751110752423603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,3584,0.018983999888102215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,16384,0.06429511308670044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,3072,0.019069333871205647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,3072,0.017474666237831116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,2560,0.0162782222032547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,2560,0.01646488904953003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,2560,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,12288,0.04983199967278374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,2048,0.013933334085676404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,2048,0.014194667339324951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,1536,0.012774222426944308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,1536,0.011668444507651858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,1536,0.012538666526476542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,1024,0.010989333192507425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,1024,0.009278222090668148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,1024,0.010095111197895473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,768,0.010139555566840703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,768,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,768,0.008707555631796518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,512,0.008757332960764566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,512,0.006443555570311016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,512,0.007890666524569193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,8192,0.03514044483502706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,256,0.00498488876554701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,256,0.007248889240953658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,128,0.00666933341158761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,128,0.004375111311674118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,128,0.006903111106819577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,64,0.00416355538699362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,64,0.007155555817815993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2560,32,0.0042408886882993905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2560,32,0.00721244431204266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,65536,0.2538444466061062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,65536,0.2749733395046658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,6144,0.02790400054719713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,65536,0.2539128992292616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,16384,0.08981777562035455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,16384,0.06221510966618856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,12288,0.05266933308707344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,12288,0.07162844472461276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,12288,0.047854221529430814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,5120,0.02481422159406874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,10240,0.0602453351020813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,10240,0.04007733199331496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,4096,0.02129511038462321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,8192,0.04577599962552389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,8192,0.03328355484538608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,7168,0.032984889215893216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,7168,0.041802667909198336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,3584,0.019699555304315355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,7168,0.03006311257680257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,3072,0.01777866648303138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,6144,0.037069333924187556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,6144,0.02680711117055681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,2048,0.014160000615649752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,5120,0.03157600098186069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,5120,0.023909333679411147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2560,256,0.007496000164084964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,4096,0.019534221953815885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,4096,0.02034133341577318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,3584,0.01982399986849891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,3584,0.018383999665578205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,3584,0.018765333626005385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,3072,0.01794577803876665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,3072,0.01740533279048072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,3072,0.017320889565679762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,2560,0.01498222185505761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,2560,0.015644444359673392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,16384,0.06679111056857638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,2048,0.011960888902346293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,2048,0.014007111390431723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,10240,0.04483200112978617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,1536,0.010023111270533668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,1536,0.012460444536474017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,1024,0.010761777559916178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,1024,0.007898666792445714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,1024,0.009762666291660732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,768,0.00920800036854214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,768,0.006783111227883234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,768,0.008719111482302347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,8192,0.03648622168434991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,512,0.005314666777849197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,512,0.00776444458299213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,6144,0.028864890336990356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,256,0.004481777962711123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,256,0.007215110792054071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,5120,0.025234666135576036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,128,0.003941333128346337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,128,0.006785777707894643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,64,0.003626666549179289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,64,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,2048,32,0.0037697777152061462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,2048,32,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,4096,0.021515554851955835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,65536,0.20662577946980795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,65536,0.23206400871276855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,16384,0.06140444676081339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,16384,0.060101330280303955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,16384,0.05823022127151489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,2560,0.016339555382728577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,2048,0.014127999544143677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,12288,0.04596799943182203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,12288,0.044046223163604736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,10240,0.03719733489884271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,10240,0.03948533203866746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,8192,0.034224887688954674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,8192,0.030258665482203167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,8192,0.032349334822760686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,1536,0.012425777812798819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,7168,0.02642400066057841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,7168,0.02975911233160231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,512,0.008295999632941352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,6144,0.023352000448438857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,6144,0.026725333597924974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,5120,0.024028443627887305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,5120,0.019810666640599568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,5120,0.023698666029506262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,256,0.006711999989218182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,4096,0.016757332616382174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,4096,0.019913777709007263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,2048,128,0.006158222340875202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,3584,0.014967110421922473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,3584,0.018739556272824604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,65536,0.2298719882965088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,3072,0.013173333472675748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,3072,0.017010667257838778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,12288,0.048355556196636625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,2560,0.011587555209795633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,2560,0.015712888704405892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,2048,0.013558222187889947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,2048,0.010033778018421596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,2048,0.013792888985739814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,10240,0.04144355654716492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,1536,0.008631111019187504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,1536,0.011659555964999728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,7168,0.03147111005253262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,1024,0.007610666255156199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,1024,0.00925422211488088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,768,0.008657777474986183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,768,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,768,0.008423111504978603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,6144,0.027056889401541814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,512,0.005174222091833751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,512,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,4096,0.020413332515292697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,256,0.0041244443919923575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,256,0.007130666739410824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,128,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,128,0.0036675557494163513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,128,0.006724444528420766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,3584,0.018966222802797954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,64,0.003478222423129611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,64,0.006858666737874349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1536,32,0.003640888879696528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1536,32,0.006894222150246303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,3072,0.017465778523021273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,65536,0.1462613344192505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,65536,0.2313786612616645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,16384,0.042287998729281955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,16384,0.05916711356904772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,2560,0.015824000040690105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,12288,0.05256266726387871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,12288,0.04564533299869961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,1536,0.012332444389661154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,10240,0.042414221498701304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,10240,0.038903110557132296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,8192,0.028431110911899145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,8192,0.03501333461867438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,8192,0.032327112224366926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,1024,0.010445333189434476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,7168,0.03129244512981839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,7168,0.02941688895225525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,6144,0.022932445009549458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,6144,0.02642755541536543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,6144,0.026553778184784785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,512,0.007688889073001013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,5120,0.02291822267903222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,5120,0.023335999912685815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,4096,0.017248888810475666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,4096,0.015056888262430826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,4096,0.019925332731670804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,3584,0.015957333975368075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,3584,0.013221333424250284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,3584,0.018409777018758986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1536,256,0.006528888725572162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,3072,0.012236444486512078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,3072,0.01700888905260298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,2560,0.013355555633703867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,2560,0.010815111299355825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,2560,0.015318221516079374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,65536,0.19555111726125082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,2048,0.00925155563486947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,2048,0.013655111193656921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,16384,0.05365777677959866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,1536,0.0075457774930530125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,1536,0.011171555353535546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,1024,0.007821333077218797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,1024,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,12288,0.04092177748680115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,1024,0.008955555657545725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,768,0.004849777867396672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,768,0.00831911133395301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,512,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,512,0.004247111164861255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,512,0.007644444704055786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,256,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,256,0.0036888888312710654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,256,0.007065777977307637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,128,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,128,0.003371555565132035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,128,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,64,0.003162666741344664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,64,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,1024,32,0.0033111110743549135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,1024,32,0.006796444455782573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,10240,0.034705777963002525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,65536,0.11557600233289932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,65536,0.22953687773810494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,16384,0.047376887665854565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,16384,0.03466399841838413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,16384,0.0583475563261244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,7168,0.02566577825281355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,12288,0.02811466654141744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,12288,0.04500000013245476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,5120,0.02030666669209798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,10240,0.024098666177855596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,10240,0.038667556312349104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,3072,0.014715555641386243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,8192,0.01980177727010515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,8192,0.03213511241806878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,7168,0.024132443798912898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,7168,0.0177511109246148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,7168,0.029148442877663508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,6144,0.021313778228229944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,6144,0.015769778026474845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,6144,0.026193777720133465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,5120,0.01890222231547038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,5120,0.013620444469981723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,5120,0.02316444449954563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,2048,0.01167377746767468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,4096,0.01089599976936976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,4096,0.01964977714750502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,1536,0.00977955593003167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,3584,0.010060444474220276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,3584,0.01812533371978336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,1024,768,0.007057777709431118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,3072,0.009233777721722921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,3072,0.016724444097942777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,2560,0.012714666624863943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,2560,0.01181422256761127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,65536,0.17631733417510986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,2560,0.015119110544522604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,2048,0.010175111393133799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,2048,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,2048,0.013435555828942193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,1536,0.008960000342792934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,1536,0.0081404447555542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,1536,0.010592000352011787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,1024,0.007683555285135905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,1024,0.006079111248254776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,1024,0.008880889250172509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,12288,0.03660444418589274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,768,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,768,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,512,0.0063928887248039246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,512,0.004449777719047335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,512,0.007706666986147563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,256,0.0037137779096762338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,256,0.005719111197524601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,256,0.007056000332037608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,128,0.0033297776761982176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,128,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,64,0.003043555551105075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,64,0.006761777732107375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,768,32,0.00309955556359556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,10240,0.03152533372243246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,768,32,0.006793777975771163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,65536,0.08059822188483344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,65536,0.15891467200385198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,16384,0.04336000151104397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,16384,0.027463111612531874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,65536,0.22811465793185762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,16384,0.058155556519826256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,12288,0.022416889667510986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,12288,0.04467822114626566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,8192,0.026479111777411565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,10240,0.01994222237004174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,10240,0.0385155545340644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,4096,0.016216889023780823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,8192,0.016546666622161865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,8192,0.03215822246339586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,7168,0.02195022172398037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,3584,0.015125332607163323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,7168,0.014916444818178812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,7168,0.029150221082899306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,6144,0.013379555609491138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,6144,0.026131555438041687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,3072,0.013853333062595792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,5120,0.011847111086050669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,5120,0.02299555473857456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,768,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,4096,0.010043555663691627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,4096,0.019388443893856473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,768,128,0.005411555369695027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,3584,0.009400889277458191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,3584,0.018175111876593698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,12288,0.0337137778600057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,3072,0.008563555777072906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,3072,0.016562667157914903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,10240,0.028712888558705647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,2560,0.01143022212717268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,2560,0.015080889066060385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,2560,0.008198222352398766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,2048,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,2048,0.012645333177513547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,8192,0.024216888679398432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,1536,0.005469333380460739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,1536,0.010844444235165914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,6144,0.019448889626397025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,1024,0.004616888860861461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,5120,0.017328000730938382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,1024,0.008777778181764815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,768,0.0041786668201287585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,768,0.008208889100286696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,512,0.006056889063782162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,512,0.0037600000699361167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,512,0.0074986666440963745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,4096,0.01477777792347802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,256,0.0033679999825027254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,256,0.0069217777086628815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,128,0.005251555393139522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,128,0.0031155554784668815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,128,0.006523555351628198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,64,0.002941333378354708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,64,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,512,32,0.003088888815707631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,512,32,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,3584,0.013776000175211163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,65536,0.05593422386381361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,65536,0.22754221492343477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,3072,0.012799999780125089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,16384,0.019440889358520508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,16384,0.05817244450251261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,12288,0.027943111128277246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,12288,0.017151110702090792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,12288,0.044680890109803945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,2048,0.009709333380063375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,1536,0.008228444390826756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,10240,0.015206222732861837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,10240,0.03860444492763943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,8192,0.02033155494266086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,8192,0.01329422245422999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,8192,0.032229334115982056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,1024,0.007244444555706448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,7168,0.012491555677519904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,7168,0.029177778297000464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,6144,0.01661422186427646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,6144,0.012245333029164208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,6144,0.026014222039116755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,5120,0.014951111541854011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,5120,0.010919999745157031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,5120,0.023183999790085688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,4096,0.012964444855848948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,4096,0.009363555245929295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,4096,0.019479999939600628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,3584,0.012044444680213928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,3584,0.008707555631796518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,3584,0.018033777674039204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,3072,0.010939555863539377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,3072,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,3072,0.01648177703221639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,256,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,512,768,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,2560,0.0074551113777690465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,2560,0.014891554911931356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,2048,0.006436444405052397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,2048,0.01238222254647149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,1536,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,1536,0.01014755583471722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,1024,0.006526222245560751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,1024,0.004355555607212914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,1024,0.008781332936551835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,65536,0.1273626618915134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,768,0.004045333299371931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,768,0.00816000004609426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,16384,0.0355422231886122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,512,0.00361333332127995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,512,0.0074879998962084455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,10240,0.02405511173937056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,256,0.003245333416594399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,256,0.0068862222962909276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,7168,0.018565333551830716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,128,0.002966222249799304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,128,0.006481777876615524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,64,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,64,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,256,32,0.0028808888875775864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,256,32,0.006597333484225803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,2560,0.009655111365848118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,65536,0.046367999580171376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,65536,0.22936177253723145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,2048,0.008467555873923832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,768,0.00609777785009808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,16384,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,16384,0.05766489108403524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,1536,0.0074879998962084455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,12288,0.01367288910680347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,12288,0.04496711161401537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,10240,0.02363377809524536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,10240,0.012700444294346703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,10240,0.038973334762785174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,8192,0.011162666810883416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,8192,0.032250665956073336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,512,0.005745777653323279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,7168,0.010613333019945355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,7168,0.029050668080647785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,128,0.004966222163703707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,256,256,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,6144,0.009927999642160203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,6144,0.026175111532211304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,5120,0.014644443988800049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,5120,0.009732444253232744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,5120,0.023169777459568445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,4096,0.0125937776433097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,4096,0.008960888617568547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,4096,0.019488889310095046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,3584,0.011524444652928246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,3584,0.008252444366614023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,3584,0.018180444836616516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,3072,0.007621333003044128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,3072,0.016422222057978313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,65536,0.126035557852851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,2560,0.006187555690606435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,2560,0.01421333352724711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,16384,0.03485155436727736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,2048,0.005519999812046687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,2048,0.011513777905040316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,1536,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,1536,0.00480355570713679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,1536,0.010106667048401302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,12288,0.02737066646416982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,7168,0.018234666850831773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,1024,0.004080888711743885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,1024,0.008714666797055138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,768,0.006055999961164262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,768,0.003744889050722122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,768,0.00814488861295912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,512,0.0034213333080212274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,512,0.007445333732499017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,8192,0.020111110475328233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,6144,0.016275554895401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,256,0.003100444459252887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,256,0.006851555572615729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,128,0.005045333256324132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,128,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,128,0.006498666687144174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,64,0.0027137777457634607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,64,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,128,32,0.0027395555128653846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,128,32,0.00654755574133661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,65536,0.04598666561974419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,65536,0.22868532604641387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,16384,0.014913777510325113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,16384,0.05734755595525106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,12288,0.012705778082211813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,12288,0.044687112172444664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,10240,0.011653333074516721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,10240,0.03873511155446371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,8192,0.010454222559928894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,8192,0.03198400139808655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,7168,0.009933333430025313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,7168,0.028819554381900366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,6144,0.009200000100665623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,6144,0.025941333836979334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,5120,0.00871733327706655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,5120,0.02295111119747162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,4096,0.007898666792445714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,4096,0.019299555155966017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,3584,0.007602666815121968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,3584,0.017865777015686035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,3072,0.0069013333155049225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,3072,0.0163128889269299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,2560,0.006172444257471297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,2048,0.00544088871942626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,2560,0.013939556148317126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,2048,0.01148799972401725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,1536,0.0046880000995265115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,1536,0.010127999716334874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,1024,0.004072000169091755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,1024,0.008769777913888296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,768,0.003742222156789568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,768,0.008231999973456064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,512,0.0034311111602518293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,512,0.0074577778577804565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,256,0.003050666716363695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,256,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,128,0.002874666617976295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,128,0.0064995553758409284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,64,0.002752888947725296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,64,32,0.0027724444452259275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,64,0.006554666492674086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,64,32,0.006506666541099548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,65536,0.043024000194337636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,16384,0.014290667242474027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,16384,0.057392001152038574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,12288,0.012283555335468717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,65536,0.2285777727762858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,12288,0.04498399959670173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,10240,0.01124622258875105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,10240,0.03891377647717794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,8192,0.00999111102686988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,7168,0.009303111169073317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,8192,0.03208622336387634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,6144,0.00869422240389718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,7168,0.028763555818133887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,5120,0.00849955528974533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,6144,0.025987555583318073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,4096,0.007599111232492659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,5120,0.022841778066423204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,3584,0.0075333333677715724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,4096,0.01941688855489095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,3072,0.006681777950790193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,3584,0.01790844400723775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,2560,0.005995555470387141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,3072,0.01633066601223416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,2048,0.005351999981535806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,2560,0.013849777479966482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,1536,0.004648000001907349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,2048,0.01149866647190518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,1024,0.004020444634887907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,1536,0.009979555176364051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,768,0.003630222131808599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,1024,0.008707555631796518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,512,0.003318222239613533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,768,0.008072888685597314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,256,0.0029680000411139596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,512,0.007416889071464539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,256,0.006807111203670502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,128,0.0027928888383838865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,128,0.006426666759782367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,64,0.002607999990383784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,64,0.006504000061088138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,768,32,32,0.0026168889469570587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,768,32,32,0.006520000182920032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,16384,1.6220053566826715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,2560,0.009536888864305284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,3072,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,16384,0.7743582195705838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,12288,1.140939606560601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,12288,0.5967217551337348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,2048,0.008420444197124904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,12288,0.5572426584031847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,10240,0.49551910824245876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,10240,0.9517644246419271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,8192,0.3907875484890408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,1024,0.006414222220579783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,10240,0.48169422149658203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,512,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,8192,0.8277262051900228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,768,128,256,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,7168,0.6859413252936469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,7168,0.3427262306213379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,8192,0.37660977575514054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,7168,0.3373120095994737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,6144,0.2988622188568115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,6144,0.6264773474799262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,6144,0.28461866908603245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,5120,0.46085866292317706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,4096,0.3743937810262044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,5120,0.23788621690538195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,4096,0.19468622737460664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,3584,0.32771465513441295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,16384,1.0054071214463975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,3584,0.18775644567277697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,3072,0.15959644317626953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,3584,0.17196355925665963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,3072,0.2849840058220757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,2560,0.2405866781870524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,3072,0.1506862243016561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,2560,0.1274053388171726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,2048,0.11337155765957302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,2048,0.20204267236921522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,1536,0.08979555633332993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,2048,0.10724000136057536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,1536,0.15015910731421575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,1536,0.08552444643444485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,1024,0.06717422273423937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,1024,0.10550044642554389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,768,0.05856977568732368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,768,0.0841146641307407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,768,0.05553955501980252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,512,0.06405066781573825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,512,0.048143111997180514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,256,0.037552889850404524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,256,0.04514044523239136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,256,0.04187466700871786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,128,0.03470755616823832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,128,0.03663555449909634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,64,0.03173866536882188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,128,0.039910223748948835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,64,0.04219555523660448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,65536,32,0.03221333358022902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,65536,32,0.04350577791531881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,65536,1.678285280863444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,65536,0.80148712793986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,65536,0.7914222081502279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,16384,0.3779546684688992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,16384,0.2138000064425998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,12288,0.15862311257256403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,12288,0.2719715436299642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,10240,0.14271199703216553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,12288,0.14873422516716853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,10240,0.22753511534796822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,10240,0.12460533777872722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,8192,0.10683466990788777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,8192,0.1831199990378486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,8192,0.10326133171717326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,7168,0.094760000705719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,7168,0.16100266244676378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,7168,0.0902151134279039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,6144,0.1389520035849677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,6144,0.08206933074527316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,5120,0.0698782205581665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,6144,0.079857779873742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,5120,0.25044798851013184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,4096,0.20823289288414848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,5120,0.11639200316535102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,4096,0.09452444314956665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,5120,0.06682222419314914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,4096,0.06119022104475233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,4096,0.053470220830705434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,3584,0.08367644415961371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,3584,0.0481333335240682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,3072,0.04529955652025011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,3072,0.07219644387563069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,3072,0.04240977764129639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,2560,0.06124622291988797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,2560,0.039814223845799766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,2560,0.037408000893063016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,2048,0.04996444450484382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,2048,0.036417778995301985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,2560,0.13889243867662218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,2048,0.031926221317715116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,1536,0.02881866693496704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,1536,0.028005331754684448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,1536,0.0410426656405131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,1024,0.022038223014937505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,1024,0.02854666776127285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,768,0.02426311042573717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,768,0.018340445227093168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,512,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,1024,0.06650044520696004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,512,0.017867555220921833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,512,0.0158595558669832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,256,0.014173333843549093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,256,0.012343111137549082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,256,0.014209777116775513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,128,0.010118222071064843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,128,0.012042666474978128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,64,0.008183110919263627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,64,0.012647111382749347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,16384,32,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,16384,32,0.013416888813177744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,65536,512,0.048172444105148315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,65536,1.194230185614692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,65536,0.6153368949890137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,16384,0.16131911012861463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,16384,0.28257155418395996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,16384,0.14923289087083605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,12288,0.11760355366600884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,12288,0.2124782270855374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,16384,0.20111110475328234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,65536,0.6042924457126194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,10240,0.17864355776045057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,10240,0.09936533371607463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,12288,0.11412978172302246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,8192,0.08243378003438313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,8192,0.1426488823360867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,8192,0.07814577553007337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,10240,0.10225422514809503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,7168,0.12560533152686224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,7168,0.06937599844402738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,6144,0.1085217793782552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,6144,0.060008002652062305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,5120,0.05636799997753567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,5120,0.09115022420883179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,5120,0.05414488911628723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,4096,0.04420444369316101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,4096,0.0750817788971795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,4096,0.04220799936188591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,3584,0.04124622212515937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,3584,0.06502666738298205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,3584,0.03877066572507223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,3072,0.056306666798061795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,3072,0.03459199931886461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,2560,0.031480888525644936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,2560,0.047650665044784546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,2560,0.030511998467975195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,2048,0.02699911097685496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,2048,0.03927110963397556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,2048,0.02622844444380866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,1536,0.023150222169028387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,1536,0.030584888325797185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,1536,0.02091022166940901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,1024,0.01902933253182305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,1024,0.02199111051029629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,1024,0.016857778032620747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,768,0.017287110288937885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,3584,0.052058666944503784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,768,0.017507554756270517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,512,0.014866666661368476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,768,0.01519733336236742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,512,0.013397333522637686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,512,0.013716444373130798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,256,0.009885333478450775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,256,0.0125591109196345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,256,0.011015999648306104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,128,0.010948444406191507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,128,0.00776533368561003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,64,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,128,0.009705777797434065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,64,0.010255999863147736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,12288,32,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,12288,32,0.010738666686746808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,768,0.02017333275742001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,1024,0.022450667288568284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,16384,128,0.012445333103338877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,16384,0.22790310117933485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,65536,0.9169457753499349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,16384,0.14334844218360052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,12288,0.17306221856011283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,65536,0.5168115297953287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,12288,0.11224355962541367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,10240,0.09113866753048366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,65536,0.6031386587354872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,10240,0.14534666803148058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,8192,0.07157511181301542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,10240,0.10561155610614353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,8192,0.11692799462212457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,8192,0.07722222142749362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,7168,0.10338844193352593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,7168,0.06681599881913927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,6144,0.08925155798594157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,6144,0.057709336280822754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,5120,0.04738044407632616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,5120,0.0750453339682685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,5120,0.04942399925655789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,4096,0.039623111486434937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,4096,0.06060889032151964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,4096,0.04106222258673774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,7168,0.07144711414972942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,6144,0.06193955739339193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,3584,0.0358497765329149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,3584,0.05417155557208591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,3584,0.037808001041412354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,3072,0.031480001078711614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,3072,0.04720266660054525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,3072,0.03300177719857957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,2560,0.028027554353078205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,2560,0.0396071109506819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,2560,0.02888710962401496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,2048,0.02427022159099579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,2048,0.0329422222243415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,2048,0.024341333243581984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,1536,0.02045866681469811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,1536,0.026565333207448322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,1536,0.019890666007995605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,12288,3072,0.035113778379228376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,1024,0.016878222425778706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,1024,0.018575999471876357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,1024,0.01664177742269304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,768,0.015264888604482016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,768,0.015448000695970325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,768,0.015046222342385186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,512,0.01296533313062456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,512,0.013302221894264221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,512,0.012554667062229581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,256,0.010942222343550788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,256,0.009204444785912832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,256,0.01016533292002148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,128,0.006990222467316522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,128,0.00980177770058314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,128,0.009363555245929295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,64,0.006172444257471297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,10240,32,0.0064195555945237475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,64,0.00996800015370051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,10240,32,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,16384,0.10762400097317165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,16384,0.17995911174350313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,65536,0.7147662374708387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,16384,0.10353777805964152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,65536,0.4078666634029812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,12288,0.13645866182115343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,10240,0.07027822070651583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,12288,0.09295644362767537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,10240,0.11431645022498237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,10240,0.07029333379533556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,8192,0.057977775732676186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,8192,0.09275466865963405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,8192,0.054809778928756714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,7168,0.055255111720826894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,7168,0.08151377571953668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,7168,0.05196088883611891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,16384,0.13235555754767522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,6144,0.07124355766508314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,5120,0.04148799843258328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,6144,0.04236800140804715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,5120,0.058913780583275684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,5120,0.036729776197009616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,4096,0.03352088729540507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,4096,0.04731200138727824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,12288,0.10365511311425103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,4096,0.031312889522976346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,3584,0.029818667305840388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,3584,0.04281066523657905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,3584,0.028784887658225164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,3072,0.03715822100639343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,3072,0.025417778227064345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,2560,0.03152088986502753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,2560,0.02234577801492479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,2048,0.021623111433453027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,2048,0.025900445050663416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,2048,0.01885600056913164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,1536,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,1536,0.020524443851576913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,1536,0.01608622239695655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,7168,0.06509688827726576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,1024,0.015740444262822468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,1024,0.013353777428468069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,10240,6144,0.055120001236597695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,768,0.01292266696691513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,768,0.012341332932313284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,65536,0.40800799263848203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,512,0.01254666679435306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,512,0.013033777475357056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,12288,0.08259466621610853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,256,0.00777599960565567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,256,0.008721777962313758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,128,0.009512888888518015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,128,0.006151999864313338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,128,0.008195555872387355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,64,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,64,0.008529778156015608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,8192,32,0.005708444449636672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,8192,32,0.008775999976529015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,6144,0.0449795557392968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,2560,0.024273778001467388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,3072,0.026612444056404963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,65536,0.6609173350863987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,16384,0.19129955768585205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,1024,0.015374221735530429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,768,0.014158222410413953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,512,0.012316444681750404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,12288,0.08323022392061022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,16384,0.09936977757347955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,12288,0.12654399871826172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,10240,0.06798399819268121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,12288,0.07680710819032457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,65536,0.4006693363189697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,10240,0.10622311300701565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,10240,0.06430044439103869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,8192,0.08679022391637166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,8192,0.05273511012395223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,7168,0.07619200150171916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,7168,0.05000799894332886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,7168,0.0469706654548645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,6144,0.06494310829374525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,6144,0.0410151117377811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,5120,0.038005332152048744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,5120,0.05503288904825846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,5120,0.035659554931852556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,4096,0.04397155510054695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,4096,0.030270222160551283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,3584,0.029543111721674602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,3584,0.0377075539694892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,3584,0.027149332894219294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,3072,0.032878223392698504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,3072,0.02460355559984843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,2560,0.02436888880199856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,2560,0.02790044413672553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,2560,0.021789333886570398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,8192,256,0.010688888529936472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,2048,0.022870222727457683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,2048,0.01850133306450314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,65536,0.4164444340599908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,1536,0.02011022302839491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,1536,0.018025777406162687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,1024,0.017272000511487324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,1024,0.01460533340771993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,1024,0.014553777045673795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,16384,0.10464977555804783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,768,0.010748444332016839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,768,0.012113778127564324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,8192,0.05769155422846476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,512,0.00850044439236323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,512,0.010246222217877706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,6144,0.0447982218530443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,256,0.006288888967699475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,256,0.008591111335489485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,4096,0.032474666833877563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,128,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,128,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,64,0.004931555440028508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,64,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,7168,32,0.005197333378924264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,7168,32,0.008573333422342936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,3072,0.026621333426899377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,65536,0.5610124270121256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,2048,0.021195555726687115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,16384,0.0828311112191942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,16384,0.1403475602467855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,16384,0.10217599736319648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,65536,0.41708710458543563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,12288,0.10607022047042847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,12288,0.08103733592563205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,1536,0.018400889303949144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,10240,0.08913866678873698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,10240,0.062863998942905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,8192,0.04420888755056593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,8192,0.07201955715815227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,8192,0.05115555392371284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,768,0.014447111222479077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,7168,0.0633937782711453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,7168,0.046817776229646474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,512,0.012568000290128918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,6144,0.0352168877919515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,6144,0.053802665736940175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,6144,0.040758222341537476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,5120,0.04612444506751167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,5120,0.035283555587132774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,4096,0.026165333059098985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,4096,0.037428445286220975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,256,0.010642666783597736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,4096,0.029548443026012842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,3584,0.03327466713057624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,3584,0.026770666241645813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,7168,128,0.009862222605281407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,3072,0.021591110361946955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,3072,0.028909332222408716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,3072,0.023864888482623633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,2560,0.019924443629052904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,2560,0.02474400069978502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,2560,0.021366222037209406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,2048,0.017498667041460674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,2048,0.02051377793153127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,2048,0.018191999859280057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,1536,0.01630577776167128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,1536,0.015582222077581616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,1024,0.013060444758998023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,1024,0.012345777617560493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,65536,0.31607643763224286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,1024,0.013202666408485837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,12288,0.06334133280648126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,768,0.010553778045707278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,768,0.011959111524952782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,512,0.01129155523247189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,512,0.008565333154466417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,512,0.009626666704813639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,256,0.006155555446942647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,256,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,10240,0.05372622278001574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,128,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,128,0.007828444242477417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,64,0.004664888812435998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,64,0.008339555727110969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,6144,32,0.00489777781897121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,6144,32,0.008647999829716152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,65536,0.27405333518981934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,7168,0.040031111902660794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,65536,0.46648799048529727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,16384,0.07825333542293973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,65536,0.38350489404466415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,16384,0.12359378072950576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,12288,0.059537774986690946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,16384,0.09884711106618245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,12288,0.09080711338255142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,12288,0.07341600126690336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,10240,0.07579644521077473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,10240,0.06195733282301161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,5120,0.030264887544843886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,8192,0.060524443785349526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,8192,0.050808889998330005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,3584,0.024056888288921777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,7168,0.05599911345375908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,7168,0.045887112617492676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,1536,0.015189333094490899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,6144,0.04695555567741394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,6144,0.04049244523048401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,5120,0.027698665857315063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,5120,0.03975555631849501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,5120,0.034909334447648786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,4096,0.023765333824687537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,4096,0.032553778754340276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,4096,0.029112888707054987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,768,0.012335110869672565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,3584,0.02882933285501268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,3584,0.026672000686327618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,128,0.008736888567606607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,6144,256,0.00979200005531311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,3072,0.02529688841766781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,3072,0.023914666639433965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,2560,0.017844445175594754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,2560,0.021851556168662176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,2560,0.020974222156736586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,2048,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,2048,0.01817599932352702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,2048,0.018419555491871305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,1536,0.013859555953078799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,1536,0.015108444624476962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,1536,0.01570755574438307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,1024,0.012015110916561551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,1024,0.01163111130396525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,1024,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,768,0.011024889018800525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,768,0.010346666806273991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,768,0.011317333413494958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,512,0.008413333031866286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,512,0.010103999740547603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,512,0.009397333694828881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,256,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,256,0.00833777752187517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,10240,0.04842577709092034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,128,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,128,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,64,0.004655999855862724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,64,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,5120,32,0.004865777575307422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,8192,0.0399635566605462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,5120,32,0.008394666843944127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,65536,0.36826221148173016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,65536,0.23715644412570527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,7168,0.03597866826587253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,6144,0.03198400139808655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,16384,0.09257244401507908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,65536,0.2381840017106798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,16384,0.058913780583275684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,12288,0.07104621993170844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,12288,0.04580355683962504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,10240,0.041209777196248375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,10240,0.062310225433773465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,10240,0.03968355390760634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,8192,0.047704888714684375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,3584,0.021899556120236714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,8192,0.03266755408710904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,7168,0.04208622376124064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,7168,0.030898667044109766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,7168,0.02983733349376255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,6144,0.03654044535424974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,6144,0.026541334059503343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,5120,0.023880889018376667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,3072,0.019661333825853135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,5120,0.030834668212466772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,5120,0.02313066687848833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,128,0.007818666597207388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,5120,256,0.008610666626029545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,4096,0.025356445047590468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,4096,0.019883554842736986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,3584,0.01926399932967292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,3584,0.022750222020679053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,3072,0.01998222205373976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,3584,0.018547554810841877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,3072,0.017103999853134155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,3072,0.016872889465755887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,2560,0.01703999936580658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,2560,0.01576266686121623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,2560,0.015357333752844067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,2048,0.01408266690042284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,2048,0.014261333478821648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,2048,0.013648000028398303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,1536,0.01184177812602785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,1536,0.012171555724408893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,1024,0.009197333620654212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,1024,0.011071999867757162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,1024,0.009929777847396003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,768,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,768,0.008712000317043727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,512,0.009381333159075843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,512,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,512,0.007696888513035244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,16384,0.06212977568308512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,256,0.004956444518433677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,256,0.007111999723646376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,128,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,128,0.00435644429590967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,128,0.006772444479995304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,64,0.004011555678314633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,64,0.006999111009968652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,4096,32,0.00416355538699362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,4096,32,0.007095110913117726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,12288,0.0477786660194397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,8192,0.0344106654326121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,65536,0.2457351154751248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,65536,0.31795112291971844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,6144,0.02715822226471371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,16384,0.08165155516730414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,16384,0.056518223550584584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,12288,0.06278222137027316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,65536,0.22789600160386828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,12288,0.04413777920934889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,10240,0.05284888876809014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,10240,0.03808088766203986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,8192,0.035192890299691096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,8192,0.042675554752349854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,4096,0.020586666133668687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,8192,0.032058666149775185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,7168,0.03760977917247348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,7168,0.028855111863878038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,1536,0.012637333737479316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,6144,0.032412442896101214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,768,0.010468444062603844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,6144,0.025733333494928148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,5120,0.028582221931881372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,5120,0.024812445044517517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,5120,0.022780444886949327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,4096,0.02252799934811062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,4096,0.019523556033770244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,4096,256,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,3584,0.020072888996866014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,3584,0.02019466625319587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,3584,0.018218666315078735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,3072,0.017682666579882305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,3072,0.01772533357143402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,3072,0.01661422186427646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,2560,0.01621866722901662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,2560,0.015299555328157214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,2560,0.015200000670221118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,2048,0.014320888453059726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,2048,0.012910222013791403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,2048,0.013564444250530668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,1536,0.012755555411179861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,1536,0.010811555716726514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,1536,0.012123555772834353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,1024,0.011267555256684622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,1024,0.008527999950779809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,1024,0.010266666611035665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,768,0.01012711144155926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,768,0.007479999628331926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,768,0.009130666653315226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,512,0.009154666629102495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,512,0.0057795556883017225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,512,0.007864889171388414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,256,0.0073831114504072405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,256,0.004690666579537922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,256,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,128,0.006680000159475539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,128,0.004148444367779626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,128,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,64,0.0037928890022966596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,64,0.0068853331936730286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3584,32,0.003986666599909465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3584,32,0.0069857777820693115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,65536,0.22365334298875597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,65536,0.29016534487406415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,65536,0.21708533498975965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,16384,0.060305780834621854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,16384,0.07228977812661065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,16384,0.05455466773774889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,12288,0.046923554605907865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,12288,0.054809778928756714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,12288,0.042005333635542125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,10240,0.04885511265860664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,10240,0.03728266557057699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,16384,0.06415733363893297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,8192,0.033722665574815534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,8192,0.03693155447642008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,8192,0.03110577662785848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,7168,0.03275022241804335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,7168,0.02834222382969326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,10240,0.043136000633239746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,12288,0.048715555005603366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,6144,0.02942311101489597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,6144,0.02548622257179684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,5120,0.024596444434589807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,5120,0.02264533274703556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,7168,0.03161777721510993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,4096,0.019729778170585632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,6144,0.028091556496090356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,4096,0.019384000036451552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,3584,0.017606221967273288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,3584,0.01797422269980113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3584,4096,0.0214711113108529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,3072,0.015596444408098856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,3072,0.01649688846535153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,2560,0.015288000305493673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,2560,0.013711111413107978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,2560,0.015068444940778943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,2048,0.013640888863139682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,2048,0.011491555306646558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,2048,0.013567999833159976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,1536,0.012123555772834353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,1536,0.009692444569534725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,1536,0.012109333442317115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,10240,0.04081688986884223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,1024,0.007839110990365347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,1024,0.009501333038012186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,768,0.009676444033781687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,768,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,768,0.008480889101823172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,512,0.008397333323955536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,512,0.005416888743638992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,512,0.007600000335110559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,7168,0.030292444758945044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,256,0.00443733317984475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,256,0.006711111300521427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,256,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,128,0.006212444355090459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,128,0.003964444415436851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,128,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,64,0.003656889001528422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,64,0.006807111203670502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,3072,32,0.0037848887344201407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,3072,32,0.006874666859706243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,6144,0.02647022240691715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,5120,0.023265777362717524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,65536,0.2411502202351888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,65536,0.2074986696243286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,4096,0.02017600006527371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,16384,0.061871998839908175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,65536,0.2087848848766751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,16384,0.05362311005592346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,12288,0.04943199952443441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,3584,0.018713777263959248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,10240,0.048693332407209605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,12288,0.04258844587537977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,10240,0.03661422265900506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,3072,0.016574222180578444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,8192,0.031164444155163232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,8192,0.03067022230890062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,8192,0.03176888823509216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,7168,0.03216000066863166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,7168,0.02784622377819485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,6144,0.028963555892308552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,6144,0.025067556235525344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,5120,0.025519111090236243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,5120,0.022318222456508215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,4096,0.019018666611777413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,4096,0.01755199995305803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,4096,0.019179556104871962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,3072,1024,0.01057866629627016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,3584,0.016979555288950603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,3584,0.01780177818404304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,3072,0.015662221444977652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,3072,0.01611822181277805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,3072,0.0163111107216941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,16384,0.05730399820539686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,2560,0.013620444469981723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,2560,0.014783999986118741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,10240,0.03763733307520548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,2048,0.011586666935020022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,12288,0.04417333337995741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,2048,0.013387555877367655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,1536,0.011271110839313932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,7168,0.028055111567179363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,1536,0.009331555830107795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,1536,0.011853333148691388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,1024,0.009740444521109263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,1024,0.009184888667530483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,1024,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,768,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,768,0.00851466672288047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,768,0.008164444731341468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,512,0.005009777843952179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,6144,0.024551999237802293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,512,0.0075128889746136135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,256,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,256,0.00423555572827657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,5120,0.021752888957659405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,256,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,128,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,128,0.0038008888562520346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,64,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,128,0.00655733338660664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,64,0.006703111032644908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2560,32,0.003638222399685118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2560,32,0.0067528887755341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,3584,0.017551110850440133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,65536,0.1909466716978285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,16384,0.053744888967937894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,65536,0.20459111531575522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,16384,0.054251554939481944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,12288,0.05035733183224996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,12288,0.04167644513977898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,10240,0.03312622176276313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,10240,0.04327999883227878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,10240,0.03583555420239767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,8192,0.026111111044883728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,8192,0.031381332212024264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,8192,0.029859556092156306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,7168,0.023750222391552393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,7168,0.02954044606950548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,7168,0.027131555808915034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,6144,0.0214373336897956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,6144,0.02605244517326355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,6144,0.02437155610985226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,5120,0.023166222704781428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,5120,0.021962665849261816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,2560,0.014312000738249885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,4096,0.01704266667366028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,4096,0.018930666976504855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,3584,0.015627556376987033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,3584,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,3584,0.017613333132531907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,3072,0.013935110635227628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,3072,0.014648000399271647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,3072,0.016186666157510545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,2560,0.013024888932704926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,2560,0.012542222109105853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,2560,0.014658666319317289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,2048,0.011384000380833944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,2048,0.010556444525718689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,2048,0.013215111361609565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,1536,0.010375111467308467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,1536,0.008731555607583787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,1536,0.011360888679822287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,2048,0.012797333300113678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,1024,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,1024,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2560,512,0.007224000162548489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,768,0.005440000030729506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,768,0.008103110724025303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,16384,0.045443554719289146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,512,0.0046862223082118565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,512,0.007430222299363878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,256,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,256,0.003927999900446998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,12288,0.036870221296946205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,256,0.0068604445291890045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,128,0.0035573335157500375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,128,0.005786666853560342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,128,0.006424888968467712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,65536,0.1694622172249688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,64,0.0033404444240861465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,64,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,2048,32,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,2048,32,0.0034728890491856467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,65536,0.1471102237701416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,16384,0.04977866676118639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,65536,0.20341155264112684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,16384,0.042224887344572276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,16384,0.05344088872273763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,12288,0.038035555018319026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,12288,0.0403039985232883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,12288,0.04164977868398031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,10240,0.03231199913554721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,10240,0.03755377729733785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,5120,0.019170666734377544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,10240,0.035818666219711304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,8192,0.02629599968592326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,8192,0.02644177774588267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,8192,0.030491554074817236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,7168,0.025711999999152288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,7168,0.02420266634888119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,7168,0.027228444814682007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,6144,0.021535999245113794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,6144,0.02197422252760993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,6144,0.02450399928622776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,5120,0.018814222680197824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,5120,0.022103110949198406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,4096,0.016572443975342643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,4096,0.014946666028764514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,4096,0.019073777728610568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,4096,0.016885333591037326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,3584,0.013956444131003486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,3584,0.017460443907313876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,1024,0.0084906667470932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,3072,0.012402666939629449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,3072,0.01611199975013733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,2560,0.01257066677014033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,2560,0.010809777511490716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,2560,0.01461866663561927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,2048,0.009213333328564962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,2048,0.013027555412716336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,1536,0.009617778162161509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,1536,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,1536,0.01107377807299296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,1024,0.007350222104125553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,1024,0.00612088872326745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,1024,0.00869066682126787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,768,0.0075706665714581805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,768,0.004958222309748332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,2048,512,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,768,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,512,0.004173333446184794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,512,0.00739644467830658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,65536,0.17914311091105142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,256,0.003649777836269803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,256,0.006872889068391588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,5120,0.01923733287387424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,128,0.003321777822242843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,128,0.006485333459244833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,64,0.0031146665828095544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,64,0.006575111299753189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1536,32,0.0032666667054096856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1536,32,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,65536,0.13257155153486463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,3584,0.01536622146765391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,65536,0.10091911421881782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,65536,0.20305244127909342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,16384,0.03173066510094537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,3072,0.013989332649442883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,16384,0.05323733223809136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,12288,0.029912889003753662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,12288,0.04128444525930617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,10240,0.025711110896534387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,10240,0.02988977895842658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,10240,0.03589422172970242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,2048,0.011335999601417117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,8192,0.021512000097168818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,8192,0.030216889248953924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,7168,0.01945688823858897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,7168,0.020240000552601285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,7168,0.02735022207101186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,768,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,6144,0.01791466606987847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,6144,0.01610933409796821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,6144,0.024474667178259954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,5120,0.01572355628013611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,5120,0.014380445082982382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,5120,0.021910222040282354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,4096,0.01389244364367591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,4096,0.011591999895042844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,4096,0.018711111611790128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,3584,0.012791111237472959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,3584,0.01092266705301073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,3584,0.01742044422361586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,3072,0.010513777534166971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,3072,0.011623111036088733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,3072,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,2560,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,2560,0.010640000303586325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,2560,0.014518222875065274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,2048,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,2048,0.008926221893893348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,2048,0.012207110722859701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,1536,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,1536,0.006542222367392645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,1536,0.009888889061080085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,512,0.006339555399285422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,1024,0.004743111216359668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,256,0.005699555493063397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,1024,0.006746666712893381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,1024,0.008639111287064022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,768,0.00425866660144594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,768,0.008045333127180735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,512,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,512,0.007432888779375289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,256,0.005403555515739653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,256,0.003452444449067116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,256,0.006782222125265334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,128,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,128,0.005119110975000593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,128,0.006401777681377199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,64,0.0030177777840031516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,64,0.006506666541099548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,1024,32,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,1024,32,0.006532444308201472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,65536,0.1209297842449612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,65536,0.0794853303167555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1536,128,0.005405333307054307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,16384,0.033957332372665405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,65536,0.20156355698903403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,16384,0.026356443762779236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,16384,0.052665776676601835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,12288,0.024808888634045918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,12288,0.04107288850678338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,10240,0.02321333355373806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,10240,0.02215911116864946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,10240,0.03600000010596381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,8192,0.01959377858373854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,8192,0.016368889146380954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,8192,0.030032889710532293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,7168,0.01817244456874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,7168,0.014925332532988654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,7168,0.02715911136733161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,6144,0.016238222519556682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,6144,0.013342222405804528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,6144,0.0243520008193122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,16384,0.0366906656159295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,5120,0.014039110806253223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,5120,0.021856889128684998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,4096,0.011526222030321756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,4096,0.018564444449212816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,12288,0.029454221328099568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,3584,0.0105591110057301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,3584,0.017482666505707633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,3072,0.010694222317801582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,3072,0.009240888886981541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,3072,0.01606755620903439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,8192,0.02146311104297638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,2560,0.008219555848174626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,2560,0.013951111170980664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,512,0.005848888721730974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,2048,0.007190222541491191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,2048,0.011384000380833944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,1024,768,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,1536,0.006015111174848344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,1536,0.009927999642160203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,1024,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,1024,0.005043555465009478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,1024,0.008582221964995066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,768,0.00619911112719112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,768,0.004540444248252445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,768,0.008028444316652086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,512,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,512,0.003966222206751506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,512,0.007360888852013483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,12288,0.026613333159022864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,256,0.0032844444115956626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,256,0.006776888751321369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,128,0.005099555684460534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,128,0.003022222262289789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,128,0.0064035554726918536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,64,0.002895999906791581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,64,0.006500444478458828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,768,32,0.002965333354141977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,768,32,0.006481777876615524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,65536,0.11104089021682739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,65536,0.05823644664552477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,65536,0.20100800196329752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,16384,0.03134222163094415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,16384,0.02126933303144243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,16384,0.05220266514354282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,5120,0.014597333139843412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,12288,0.01868088874551985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,12288,0.04151733385192023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,10240,0.021877333521842957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,10240,0.017352888981501263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,10240,0.03615377677811517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,4096,0.013030221892727746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,8192,0.015617777903874716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,8192,0.03014577759636773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,3584,0.012074666718641916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,7168,0.014113777213626437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,7168,0.026897778113683064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,2048,0.008240000241332585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,6144,0.014312888185183207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,6144,0.024332443873087566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,2560,0.009516444471147325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,5120,0.013253333667914072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,5120,0.021718222233984206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,1536,0.007510222494602203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,4096,0.010297777752081553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,4096,0.018572444717089336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,768,256,0.005405333307054307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,3584,0.009157333109113906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,3584,0.017197334104114108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,3072,0.009954666925801171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,3072,0.008408889174461365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,3072,0.015507555670208402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,2560,0.008718222379684448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,12288,0.025049777494536504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,2560,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,2560,0.013438222308953604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,2048,0.006375999914275275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,2048,0.011295110815101199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,1536,0.007027555671003129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,1536,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,1536,0.009835555321640438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,1024,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,1024,0.004600888739029567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,1024,0.00853066643079122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,768,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,768,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,768,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,512,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,512,0.003649777836269803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,512,0.007381333245171442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,256,0.005137777576843898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,256,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,256,0.006711999989218182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,128,0.004877333425813251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,128,0.0029297777348094513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,128,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,64,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,64,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,512,32,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,512,32,0.006411555740568373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,65536,0.10897599988513523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,65536,0.04951999915970696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,6144,0.015071110592948066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,65536,0.20356267028384736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,16384,0.014516444669829475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,16384,0.05314133233494229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,12288,0.024495111571417913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,12288,0.012609778179062737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,12288,0.04158400164710151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,10240,0.021334222621387903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,10240,0.011332444846630096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,10240,0.0358915560775333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,8192,0.018180444836616516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,8192,0.010112889111042023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,8192,0.018717333674430847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,8192,0.030071109533309937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,7168,0.010413332945770687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,7168,0.016734222571055096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,7168,0.027085334062576294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,6144,0.012268444730175866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,6144,0.024307555622524683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,5120,0.013179555535316467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,5120,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,5120,0.021554667088720534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,4096,0.011166222393512726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,4096,0.008936000367005667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,4096,0.018241778016090393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,3584,0.010391111175219217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,3584,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,7168,0.017083555459976196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,3584,0.01632000009218852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,3072,0.009533333281675974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,3072,0.00721866637468338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,3072,0.014428445034556918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,2560,0.008488888541857401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,2560,0.0063742221229606206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,2560,0.012412444584899478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,2048,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,2048,0.011005333728260465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,1536,0.0068417779273457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,1536,0.00490488898422983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,1536,0.009745777481132084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,1024,0.006142222219043308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,1024,0.0041511112617121804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,1024,0.008486222061845992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,768,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,768,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,768,0.007890666524569193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,4096,0.012160888976520963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,512,0.003644444462325838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,512,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,256,0.005180444568395615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,256,0.003168888803985384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,256,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,5120,0.01370133293999566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,128,0.002895999906791581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,128,0.006340444501903322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,64,0.0027297778675953546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,64,0.0064026667839950984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,256,32,0.0027582223216692605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,256,32,0.006396444307433234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,3584,0.011241777903503843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,65536,0.039495110511779785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,65536,0.2043422195646498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,16384,0.030164443784289893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,16384,0.013190222283204397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,16384,0.053363555007510714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,512,2048,0.007879999776681265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,12288,0.011544889046086205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,12288,0.041830221811930336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,16384,0.030649777915742662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,10240,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,10240,0.03573333223660787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,6144,0.014908444550302295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,128,0.005015111217896144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,8192,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,8192,0.030015998416476782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,2048,0.007912889122962952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,7168,0.016520000166363187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,7168,0.008825778133339351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,7168,0.027522666586769953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,6144,0.014695111248228284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,6144,0.010074666804737514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,5120,0.00923644420173433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,6144,0.02459822263982561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,5120,0.02148177723089854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,4096,0.011029333704047732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,4096,0.007861333588759104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,4096,0.01833866702185737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,3584,0.0075128889746136135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,3584,0.01629066632853614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,3072,0.009475555684831407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,3072,0.00684799998998642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,3072,0.014056000444624158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,2560,0.008524444368150499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,256,512,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,2560,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,2560,0.012461333639091916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,2048,0.007751999629868402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,2048,0.005511111269394557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,2048,0.01111822244193819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,1536,0.004825777891609403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,1536,0.009830222361617619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,1024,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,1024,0.0041200001206662925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,1024,0.008618666893906063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,768,0.003752888904677497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,768,0.00797155582242542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,512,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,512,0.003432000055909157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,512,0.007287999822033777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,256,0.005192888693677055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,256,0.0030702222138643265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,256,0.006728888799746831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,65536,0.10712711016337077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,128,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,128,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,64,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,64,0.0063884444534778595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,128,32,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,128,32,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,65536,0.03386933273739285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,65536,0.20480710930294463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,16384,0.012114666402339935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,16384,0.053195556004842125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,12288,0.010644444160991244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,12288,0.041448887851503156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,10240,0.009661333428488838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,10240,0.035806222094429865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,8192,0.008907555705971187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,8192,0.02998222245110406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,7168,0.008661333057615492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,7168,0.027066666218969557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,6144,0.007839110990365347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,6144,0.024227556255128648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,5120,0.007801777786678738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,5120,0.021568000316619873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,4096,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,4096,0.01791911158296797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,3584,0.007176888485749562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,3584,0.015655999382336933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,3072,0.006682666639486949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,3072,0.013827555709415011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,2560,0.006051555689838197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,2560,0.012359999948077731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,2048,0.005511999958091312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,2048,0.011020444333553314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,1536,0.0047120000753137796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,1536,0.009710222482681274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,1024,0.004063111212518481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,1024,0.008416000339719985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,768,0.0036275556517971885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,768,0.007840000092983246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,512,0.003297777846455574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,12288,0.024059555596775476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,512,0.007169778148333232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,256,0.002974222310715251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,256,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,128,0.002777777819169892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,128,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,64,0.002640888922744327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,64,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,64,32,0.0026364444444576898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,64,32,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,65536,0.031368000639809504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,65536,0.2035964462492201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,16384,0.011468444433477191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,12288,0.010080888867378235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,16384,0.052705778015984424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,12288,0.04127111037572225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,10240,0.009406222237481011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,8192,0.008818666968080733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,10240,0.035798221826553345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,8192,0.029967112673653498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,7168,0.008341333104504479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,7168,0.027116444375779893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,6144,0.007496889266702864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,6144,0.024273778001467388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,5120,0.007160888777838812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,5120,0.021486222743988037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,4096,0.006421333385838403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,4096,0.017795556121402316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,3584,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,3584,0.015251555376582675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,3072,0.006471111128727595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,3072,0.01369866645998425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,2560,0.006001777946949005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,2560,0.01238666640387641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,2048,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,2048,0.011056000159846412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,1536,0.0047404443224271136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,1536,0.00980355590581894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,1024,0.004078222231732475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,1024,0.008577778107590145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,768,0.0037288889288902283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,768,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,512,0.003313777761326896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,512,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,256,0.0029973333908451926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,256,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,128,0.002777777819169892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,128,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,64,0.0026177778426143858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,64,0.0063902222447925145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,512,32,32,0.0026346666531430352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,512,32,32,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,10240,0.0211191111140781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,16384,0.5936275588141547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,16384,1.252852439880371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,16384,0.5886133511861166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,12288,0.9334772957695855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,12288,0.4449333084954156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,8192,0.018066666192478605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,10240,0.3948684533437093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,10240,0.7470693588256836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,10240,0.3925911055670844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,8192,0.62918668323093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,5120,0.013031110995345645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,3584,0.01019022199842665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,8192,0.29966754383511013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,1536,0.006833777659469181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,7168,0.49263911777072483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,7168,0.2743315431806776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,512,128,128,0.005014222115278244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,6144,0.4320853286319309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,6144,0.23080711894565156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,6144,0.2371928956773546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,5120,0.3575742244720459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,4096,0.15887022018432617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,4096,0.2893448935614692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,5120,0.20006222195095488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,4096,0.16002933184305826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,3584,0.25453244315253365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,3584,0.14224355750613743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,3072,0.22117333941989473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,3072,0.12084799342685276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,2560,0.18799732791052925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,12288,0.44500891367594403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,2560,0.10468888945049709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,2048,0.1534986628426446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,2048,0.08661066823535496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,1536,0.07227644655439588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,1536,0.11994400289323594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,1536,0.07142044438256158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,1024,0.05453422334459093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,1024,0.08470488919152154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,1024,0.054437332683139376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,768,0.047565334373050265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,768,0.06854133473502265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,768,0.0458826654487186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,512,0.038409776157803006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,512,0.05141511228349474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,512,0.03994311226738824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,256,0.03064799971050686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,256,0.038218667109807335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,256,0.034019556310441755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,8192,0.30595554245842826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,128,0.025518221987618342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,128,0.03160088923242357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,128,0.035098665290408664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,64,0.026107556290096704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,64,0.03404533200793796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,65536,32,0.027614222632514104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,65536,32,0.03702755437956916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,7168,0.26719021797180176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,65536,1.2247556050618489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,5120,0.20174311267005074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,65536,0.6121342447068956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,16384,0.29974667231241864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,16384,0.15617511007520887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,3584,0.14189067151811388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,12288,0.21625778410169813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,3072,0.122071107228597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,12288,0.11977867285410564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,65536,0.6126986609564887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,2560,0.10627822081247966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,10240,0.17987288369072807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,8192,0.14485243956247965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,10240,0.10536089208390977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,8192,0.08201511038674249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,65536,2048,0.0896035565270318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,7168,0.1273395617802938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,7168,0.07279555665122138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,6144,0.06368088722229004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,6144,0.10983377695083618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,6144,0.06334133280648126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,5120,0.09286310937669541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,5120,0.05386844608518812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,4096,0.04428622126579285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,4096,0.07460533248053657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,4096,0.044459554884168834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,3584,0.040150221851136945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,3584,0.06589866346783109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,3584,0.04032444291644626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,3072,0.03566133313708835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,3072,0.05710311068428887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,3072,0.035791112316979304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,2560,0.03162044286727905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,2560,0.04873422119352552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,2560,0.03180711136923896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,2048,0.03970488905906677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,2048,0.027748444014125403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,2048,0.02750933335887061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,1536,0.0313022236029307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,1536,0.023623999622133043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,1536,0.023046221998002794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,1024,0.018940443793932598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,1024,0.022522666388087805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,1024,0.01737333337465922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,768,0.016873778568373788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,768,0.018005333013004728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,768,0.015258666541841296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,512,0.014675555957688225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,512,0.013824889229403602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,512,0.013572444518407186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,256,0.010275555153687796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,256,0.012146666646003723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,256,0.011961778004964193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,128,0.010558221903112201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,128,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,128,0.009837333526876237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,64,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,64,0.010450666977299584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,16384,32,0.006800888727108638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,16384,32,0.011394666300879585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,65536,0.8727208773295084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,65536,0.5430924627516005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,16384,0.157094226943122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,16384,0.20779289139641655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,16384,0.14179821809132895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,12288,0.10644000106387669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,16384,0.1507777770360311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,12288,0.1217413345972697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,12288,0.15609067016177708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,12288,0.11611110634273952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,65536,0.5830746756659614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,10240,0.13087556097242567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,8192,0.10548533333672418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,10240,0.0976053343878852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,8192,0.07357066869735718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,10240,0.09941511021720038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,8192,0.08331555790371366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,7168,0.09307910998662312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,7168,0.06505244308047824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,6144,0.08016710811191134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,7168,0.06783377461963229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,6144,0.0581351121266683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,6144,0.059875554508633085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,5120,0.06662755542331271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,8192,0.08034488889906141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,5120,0.052721778551737465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,4096,0.05385422375467089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,4096,0.04210933380656772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,3584,0.04802133308516609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,3584,0.03785689009560479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,3072,0.03154044349988302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,3584,0.03871466716130575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,3072,0.04239999916818407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,7168,0.07230933507283528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,2560,0.03630133469899496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,3072,0.03404533200793796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,2560,0.02996266550487942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,2048,0.0292871097723643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,2048,0.025177778469191656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,1536,0.02403555479314592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,2048,0.02534755567709605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,1536,0.020485333270496793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,1024,0.01793866687350803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,1024,0.017455110947291057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,1024,0.016872889465755887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,768,0.014496889379289416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,16384,5120,0.0539919998910692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,768,0.015993777248594496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,768,0.015152000718646579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,512,0.013529777526855469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,512,0.011529777612951068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,512,0.013398222625255585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,256,0.008791111409664154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,256,0.010586666564146677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,128,0.009822222093741098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,128,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,128,0.009461333354314169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,64,0.005678222411208683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,64,0.010095111197895473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,12288,32,0.005982222242487802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,12288,32,0.010435555544164447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,65536,0.8865928649902344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,65536,0.7095253202650281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,65536,0.4109991126590305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,16384,0.17954577340020073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,16384,0.10453333457310994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,10240,0.09154044257269965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,12288,0.14257689317067465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,12288,0.08167466852400038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,5120,0.047692444589402944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,4096,0.03987911012437608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,10240,0.1138862238989936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,10240,0.0680604444609748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,8192,0.11569421821170384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,8192,0.09355022509892781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,2560,0.028274665276209515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,8192,0.05599821938408745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,7168,0.0806808869043986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,7168,0.10284799999660915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,7168,0.04944533440801832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,1536,0.02102311121092902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,6144,0.07095644209120008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,6144,0.04354666670163473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,6144,0.09030044741100735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,5120,0.06009066767162747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,5120,0.03760977917247348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,4096,0.04890133274926079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,12288,256,0.010953777366214328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,4096,0.05896000067392985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,4096,0.03168888886769613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,3584,0.04328622089491951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,3584,0.05121866530842251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,3584,0.029233776860766943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,3072,0.03769866625467936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,3072,0.04454844527774387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,3072,0.026160889201694067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,2560,0.03737066520584954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,2560,0.03235200047492981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,2560,0.022707555029127333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,2048,0.026565333207448322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,2048,0.019141332970725167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,1536,0.021358221769332886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,1536,0.023899555206298828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,1536,0.016102222932709586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,1024,0.017775999175177682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,1024,0.01330488920211792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,1024,0.01586577792962392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,768,0.013025778035322825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,768,0.012183999849690331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,768,0.015607999430762397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,512,0.010562666588359408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,512,0.013032000097963544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,512,0.010838222172525195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,256,0.007818666597207388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,256,0.010823110739390055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,256,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,128,0.0061048890153567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,128,0.009193778038024902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,128,0.008082666330867344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,64,0.005403555515739653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,10240,32,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,32,0.009127111070685917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,10240,64,0.008551999926567078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,65536,0.5606133143107096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,65536,0.3178639941745334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,16384,0.08821155627568562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,65536,0.3974293337927924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,16384,0.1437244415283203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,16384,0.1068826648924086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,12288,0.0639466643333435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,12288,0.10898933145735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,10240,0.09137866894404094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,10240,0.054105778535207115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,12288,0.07956800195905897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,8192,0.07304088936911689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,10240,0.06808977656894259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,8192,0.055239998632007174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,7168,0.06423022349675496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,7168,0.048935112026002675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,6144,0.03587377733654446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,6144,0.05535733368661669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,6144,0.04247555467817518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,5120,0.03105688757366604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,5120,0.046572443511750966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,16384,0.22039911482069227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,5120,0.036920888556374445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,4096,0.03786755601565043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,4096,0.025511110822359722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,4096,0.03017866611480713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,3584,0.023539554741647508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,3584,0.03372800019052293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,3584,0.0281333327293396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,3072,0.021312889125612047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,3072,0.029637333419587877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,3072,0.024639111426141527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,2560,0.019273777802785236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,12288,0.16645067267947727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,2560,0.025255999631351892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,2560,0.02174844510025448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,2048,0.01719999975628323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,2048,0.02085688875781165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,2048,0.018497778309716117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,1536,0.016512889001104567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,1536,0.015771556231710646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,1024,0.012668444050682915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,1024,0.012511110968059964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,1024,0.013343999783198038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,768,0.011591110792424945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,768,0.010757333702511258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,768,0.012245333029164208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,512,0.010767999622556897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,512,0.008826666408114964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,512,0.010050666828950247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,256,0.009473777479595607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,256,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,256,0.008475555313958062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,128,0.008232889076073965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,128,0.005349333501524395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,128,0.00794222205877304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,64,0.0047208890318870544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,64,0.008346666892369589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,8192,32,0.00489155575633049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,8192,32,0.008790222307046255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,10240,0.14404088921017116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,65536,0.5098524623446995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,65536,0.880166212717692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,16384,0.12957244449191624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,5120,0.07519466347164579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,65536,0.3827013439602322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,16384,0.169466667705112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,10240,2048,0.030159999926884968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,16384,0.10175022151735093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,12288,0.09472088681326972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,10240,0.07980799674987793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,12288,0.08702133099238078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,10240,0.06518666611777411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,8192,0.06795910994211833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,8192,0.05284088850021362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,7168,0.056240889761182994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,7168,0.04782399866316053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,6144,0.04738577869203356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,6144,0.04881333311398824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,6144,0.041843556695514254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,5120,0.03885866536034478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,5120,0.041307555304633245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,5120,0.03616888986693488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,7168,0.03980088896221585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,4096,0.0340266658200158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,8192,0.04419022136264377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,4096,0.029838220940695867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,3584,0.03072711163096958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,3584,0.027719111906157598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,3072,0.02663733396265242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,3072,0.02695466743575202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,3072,0.024115555816226538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,8192,1536,0.015150222513410779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,2560,0.023207111491097346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,2560,0.021713778376579285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,2048,0.020018666982650757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,2048,0.01902755598227183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,2048,0.018304889400800068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,1536,0.016961778203646343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,1536,0.01553066737122006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,1536,0.015639111399650574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,10240,0.07834488815731473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,1024,0.012163555456532372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,12288,0.09629066785176595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,1024,0.01331999980741077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,768,0.010605333579911126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,768,0.01203555530971951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,8192,0.060531556606292725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,7168,0.05406666795412699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,512,0.0090844440791342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,512,0.009656888743241629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,256,0.009691555466916826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,256,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,256,0.008441777692900764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,128,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,128,0.007989333735571967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,64,0.005160888863934411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,64,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,7168,32,0.005428444594144821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,7168,32,0.008609777523411645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,4096,0.03215555681122674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,65536,0.2851297855377197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,65536,0.41605332162645126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,65536,0.37372978528340656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,3584,0.02980977627966139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,2560,0.023317333724763658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,16384,0.1079911126030816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,16384,0.09973422023985122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,12288,0.05653422408633762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,12288,0.08192177613576253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,12288,0.07585599687364367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,10240,0.04801244537035624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,10240,0.07035377952787611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,10240,0.06350488795174493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,8192,0.03949866692225138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,8192,0.05531466669506497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,8192,0.05134310987260607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,7168,0.03520977828237746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,7168,0.048807111051347524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,7168,0.04606399933497111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,6144,0.03148266673088074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,6144,0.042134222057130605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,6144,0.04091644287109375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,5120,0.03586133321126302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,5120,0.035190221336152815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,4096,0.023394667439990576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,4096,0.02920888861020406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,4096,0.02927466564708286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,3584,0.02181511123975118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,3584,0.02623644471168518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,3584,0.026593777868482802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,3072,0.01977422171168857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,1024,0.01428533262676663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,3072,0.022932445009549458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,3072,0.023669333921538457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,2560,0.019561777512232464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,2560,0.020994666549894545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,2048,0.016192000773217943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,2048,0.018404444058736164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,1536,0.014004444082578024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,1536,0.013564444250530668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,1536,0.01579466627703773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,1024,0.01238222254647149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,1024,0.010416000253624385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,1024,0.013188444077968597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,768,0.011324444578753578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,768,0.009117333425415887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,768,0.01166311071978675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,512,0.010247999595271217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,512,0.0075839997993575195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,512,0.009447111023796929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,256,0.008407110969225565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,256,0.005517333332035277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,256,0.008435555630260045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,128,0.007233777807818518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,128,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,128,0.007813333637184566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,64,0.0042444442709287005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,64,0.00814488861295912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,6144,32,0.00442755553457472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,6144,32,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,512,0.011541333463456897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,768,0.012986666626400419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,65536,0.4554302427503798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,7168,128,0.008159110943476358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,16384,0.09631733099619548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,65536,0.3734871016608344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,16384,0.11543467309739854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,12288,0.07056978013780382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,16384,0.058407107988993325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,12288,0.04507199923197428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,65536,0.22038755151960585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,10240,0.05952178107367622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,10240,0.04247911108864678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,8192,0.05820444557401869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,8192,0.04737155636151632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,8192,0.0328764451874627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,7168,0.042466666963365346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,7168,0.05179911189609104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,7168,0.029900444878472224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,6144,0.03693866729736328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,6144,0.02678577767478095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,5120,0.03736799955368042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,5120,0.031932443380355835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,5120,0.02355466617478265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,4096,0.03081866767671373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,4096,0.025736888249715168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,4096,0.019759999381171334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,3584,0.027651555008358423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,3584,0.023190221852726404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,3584,0.018404444058736164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,16384,0.07331733571158515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,3072,0.020277332928445604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,3072,0.016911110944218107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,2560,0.020261334048377145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,2560,0.01720977822939555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,2560,0.015304888288180033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,2048,0.01682488951418135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,2048,0.014333332578341166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,2048,0.013609777722093793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,1536,0.01425333321094513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,1536,0.01201244443655014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,1536,0.01219733307758967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,1024,0.011824888487656912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,5120,0.02759644389152527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,1024,0.009390222529570261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,1024,0.010641777680979835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,768,0.010826666322019366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,768,0.008278222547637092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,768,0.008779555559158325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,512,0.0094373333785269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,512,0.006760889043410619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,512,0.007725333174069722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,2560,0.01792711185084449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,6144,2048,0.015559110376569959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,256,0.005070222334729301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,256,0.007104000283612146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,128,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,128,0.004378666480382283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,128,0.006761777732107375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,64,0.00702400008837382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,64,0.0041831110914548235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,5120,32,0.007124444676770105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,5120,32,0.004323555363549126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,65536,0.2852444383833143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,16384,0.06039555867513021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,65536,0.2099484470155504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,16384,0.08063377936681111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,16384,0.05494844582345751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,12288,0.07125600179036458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,12288,0.043024000194337636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,10240,0.0383857786655426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,10240,0.06675110922919379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,10240,0.03719110952483283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,12288,0.08902044428719415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,8192,0.042149331834581166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,8192,0.030997332599427965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,10240,0.07548355393939547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,7168,0.028592887851927016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,7168,0.04215822286076016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,7168,0.028600888119803533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,6144,0.03896444373660617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,6144,0.025241777300834656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,6144,0.044806222120920815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,5120,0.03482755687501695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,5120,0.022312000393867493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,3072,0.023920888702074688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,4096,0.02083822256988949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,4096,0.01923199991385142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,3584,0.01752800080511305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,3584,0.01869777838389079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,3584,0.017824888229370117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,3072,0.015951111912727356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,3072,0.016605334149466622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,3072,0.016369778248998854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,5120,256,0.007548444800906711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,2560,0.013850666582584381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,2560,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,2048,0.012786666552225748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,2048,0.011757333245542316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,2048,0.013377777404255338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,65536,0.2289173338148329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,1536,0.00978133330742518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,1536,0.011966221862369113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,12288,0.04509688748253716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,1024,0.008033778104517195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,1024,0.00978311151266098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,768,0.008847999903890822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,768,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,768,0.00850399997499254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,512,0.0075093333919843035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,512,0.007583111524581909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,256,0.006158222340875202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,256,0.004455111092991299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,256,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,128,0.005392000079154968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,128,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,128,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,64,0.0035955554081334006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,64,0.006796444455782573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,8192,0.03161866797341241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,4096,32,0.003823111040724648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,4096,32,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,65536,0.2539031240675184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,65536,0.206587553024292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,16384,0.05766399701436361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,6144,0.025261332591374714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,16384,0.06932444704903497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,12288,0.04510044389300876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,16384,0.052432888083987765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,5120,0.02235466738541921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,12288,0.05312622255749173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,12288,0.04115644428465102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,10240,0.0444471107588874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,10240,0.03568089008331299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,8192,0.0374311109383901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,8192,0.029857777886920508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,7168,0.02769333289729224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,7168,0.03369866808255514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,4096,0.018739556272824604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,7168,0.027168888184759352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,6144,0.029205332199732464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,6144,0.02492888934082455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,6144,0.024480000138282776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,2560,0.014375999569892883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,5120,0.025015999873479206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,5120,0.02176800039079454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,4096,0.018208000395033095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,4096,0.020774222082561918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,4096,0.018809777167108323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,3584,0.016925333274735343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,3584,0.018439999885029264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,3584,0.017553778158293832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,3072,0.015549333559142219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,3072,0.01590666671593984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,3072,0.016107555892732408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,2560,0.014127110441525778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,2560,0.014480888843536377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,2560,0.01464355488618215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,1536,0.011366222467687396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,2048,0.012826667063766055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,2048,0.013220444321632385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,1536,0.010407110883129967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,1536,0.011899555722872416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,4096,1024,0.009857777920034196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,1024,0.008857777549160851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,1024,0.009285333255926767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,65536,0.20651377571953666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,768,0.007885333564546373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,768,0.008413333031866286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,512,0.006888888776302338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,512,0.006154666758245892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,512,0.007681777907742395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,10240,0.037436445554097496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,256,0.004595555365085602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,256,0.006859555426571105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,8192,0.030778666337331135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,128,0.004050666673315896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,128,0.006565333240562015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,64,0.0037475555307335327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,64,0.006735111276308696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3584,32,0.003927111211750242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3584,32,0.006825777805513806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,5120,0.021722666091389124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,65536,0.19345066282484266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,65536,0.22434488932291666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,65536,0.19525599479675293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,16384,0.06124889188342624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,16384,0.051222221718894116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,12288,0.04020177655749851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,12288,0.045850667688581676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,12288,0.04088977641529507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,10240,0.03474666674931844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,10240,0.0381644434399075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,10240,0.03531466590033637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,2048,0.012518222133318583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,8192,0.031256887647840716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,8192,0.02940711047914293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,1536,0.01127199994193183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,7168,0.026194666822751362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,7168,0.02792888879776001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,7168,0.02686400049262577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,6144,0.025006221400366888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,6144,0.02362222141689724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,6144,0.02418577836619483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,5120,0.020920889245139226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,5120,0.0214355554845598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,5120,0.02183288832505544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,4096,0.017489777670966256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,4096,0.01705511079894172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,4096,0.018598222070270114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,1024,0.009604444106419882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,3584,0.015415110521846347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,3584,0.017503110898865592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,3072,0.014926221635606555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,3072,0.01443377799457974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,3072,0.0158933334880405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,2560,0.011821333732869891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,2560,0.013644444445768992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,2560,0.014490667316648694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,2048,0.012352888782819113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,2048,0.011230222052998014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,2048,0.013237333132161034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,1536,0.010741333166758219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,1536,0.008797333472304875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,1536,0.011456888582971362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,1024,0.007526222202512953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,1024,0.008947555389669206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,768,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,768,0.006175999840100606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,768,0.008100444244013892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,512,0.00656088896923595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,512,0.004933333231343163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,512,0.007440889047251807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,256,0.005688888745175467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,256,0.0041351111398802865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,256,0.006888000087605582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,768,0.008450667063395182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,128,0.003671110918124517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,128,0.006522666662931442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,64,0.0034906665484110513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,64,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,3072,32,0.0036595554815398324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,256,0.005688000056478713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,3072,32,0.006836444553401735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3584,128,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,65536,0.1889493332968818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,65536,0.19283821847703722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,16384,0.05510222249560886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,16384,0.05089333322313097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,12288,0.03272355596224467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,12288,0.040179553959104754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,12288,0.040087110466427274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,16384,0.05230489042070177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,10240,0.034084445900387235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,10240,0.03475555446412828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,8192,0.029391109943389893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,8192,0.02776266634464264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,8192,0.029686222473780315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,7168,0.02180533276663886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,7168,0.024588444166713293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,7168,0.026920888159010146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,6144,0.019595555133289762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,6144,0.02144977781507704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,6144,0.02424533334043291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,5120,0.017131555411550734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,5120,0.018295110927687753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,5120,0.021340444684028625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,3584,0.01626311077011956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,4096,0.014653333359294467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,4096,0.014640889234013028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,4096,0.018375999397701688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,3584,0.014001778430408902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,3584,0.013946667313575745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,3584,0.017121778594122995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,3072,0.011882666912343768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,3072,0.015761777758598328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,2560,0.011632888681358762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,2560,0.010470222267839642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,2560,0.014281777871979607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,2048,0.010609777437316047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,2048,0.00906488878859414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,2048,0.012904889053768583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,1536,0.009481777747472128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,1536,0.007782222496138678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,1536,0.011206222077210745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,1024,0.009653333160612319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,1024,0.007342221836249034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,1024,0.00647199981742435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,1024,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,768,0.0053137776752312975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,768,0.008081778056091731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,512,0.005852444304360284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,512,0.004514666481150521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,512,0.00742222203148736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,256,0.0052168890833854675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,256,0.003926222109132343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,3072,128,0.005346666607591841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,256,0.006830222076839871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,128,0.004931555440028508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,128,0.00350222239891688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,128,0.0064737776087390045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,64,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,64,0.0033760000434186724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2560,32,0.006673777682913675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2560,32,0.0034266666819651923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,65536,0.15081866582234701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,65536,0.19045866860283744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,16384,0.04579022195604113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,16384,0.049550222025977246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,16384,0.05060088965627882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,12288,0.03471111257870992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,12288,0.047616001632478505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,12288,0.039919998910692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,10240,0.02983199887805515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,10240,0.04087022278043959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,10240,0.03418933351834615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,8192,0.024328889118300542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,8192,0.031606223848130964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,8192,0.028688000308142766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,7168,0.02215733296341366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,7168,0.028763555818133887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,7168,0.026153778036435444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,6144,0.01998044384850396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,6144,0.02476355599032508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,6144,0.023564444647894964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,5120,0.017588444881969027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,5120,0.02146666745344798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,5120,0.020956445071432326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,4096,0.014808000789748298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,4096,0.014827556080288358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,4096,0.01808888879087236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,3584,0.013803555733627744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,3584,0.013783999615245394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,3584,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,65536,0.1559644407696194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,3072,0.012878222597969903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,3072,0.015806222955385845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,16384,0.04178933302561442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,2560,0.011595555477672152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,2560,0.010713777608341642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,2560,0.01443822185198466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,2048,0.00925866680012809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,10240,0.028519110547171697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,2048,0.012692444854312472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,8192,0.02349600030316247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,1536,0.007703999678293864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,1536,0.01036800030204985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,1024,0.006555555595291986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,1024,0.005676444619894028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,1024,0.008585777547624376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,768,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,768,0.004927999857399198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,768,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,512,0.0042444442709287005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,512,0.007415999968846639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,3072,0.012646222280131446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,256,0.0036888888312710654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,256,0.006794666664467917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,128,0.0046977777447965415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,128,0.0033564445459180405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,128,0.006414222220579783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,64,0.0032248888164758682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,64,0.006545777950021956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,2048,32,0.0033457777980301115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,2048,32,0.006551110910044775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2560,768,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,65536,0.12782843907674155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,65536,0.19280978043874106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,16384,0.036638223462634616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,16384,0.036481777826944985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,65536,0.16708710458543566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,12288,0.028982222080230713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,16384,0.05059555504057142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,12288,0.03050222330623203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,12288,0.03959200117323134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,10240,0.02527644402450985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,10240,0.0262435558769438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,10240,0.03491200009981791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,8192,0.02142133315404256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,8192,0.029367112451129492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,7168,0.019162666466501024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,7168,0.01924888955222236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,7168,0.026456000076399908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,6144,0.017242666747834947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,6144,0.016734222571055096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,6144,0.023747555083698694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,5120,0.015284443895022074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,5120,0.014523555835088095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,5120,0.021223111285103693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,4096,0.012904889053768583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,4096,0.012121777567598553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,4096,0.01830044388771057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,3584,0.012302222351233164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,3584,0.010753778119881948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,3584,0.01700444519519806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,3072,0.011215999722480774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,3072,0.009543110926946005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,3072,0.015609777635998197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,2560,0.010394666757848527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,2560,0.010576888918876648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,2560,0.01421866648726993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,2048,0.008721777962313758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,2048,0.00884622252649731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,2048,0.011967110964987012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,1536,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,1536,0.007460444337791867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,1536,0.010451555252075195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,1024,0.006612444503439798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,1024,0.00545866663257281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,1024,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,768,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,768,0.004772444566090901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,768,0.008105778031879002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,3072,0.012822222378518848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,512,0.004190222256713443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,512,0.007448888487286038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,2048,0.010278222461541494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,256,0.0036026665733920205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,256,0.006868444383144379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,1536,0.008524444368150499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,128,0.00332266671790017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,128,0.0064479998416370815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,64,0.003022222262289789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,64,0.006814222368929122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1536,32,0.003069333318206999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1536,32,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,65536,0.13169866138034397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,512,0.005496888938877318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,65536,0.10023999876446194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,65536,0.1898124482896593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,16384,0.029289778735902574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,16384,0.05013244350751241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,2048,256,0.004963555683692296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,12288,0.027895109521018133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,12288,0.039701332648595176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,10240,0.02091466718249851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,10240,0.02457422183619605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,10240,0.03457866774664985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,8192,0.017502221796247695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,8192,0.017128888103697035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,8192,0.028930667373869155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,7168,0.01624000072479248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,65536,0.15707643826802573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,7168,0.01533600025706821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,7168,0.02641333308484819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,6144,0.014676445060306124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,6144,0.013627555635240344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,6144,0.02371999952528212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,5120,0.013185777597957186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,5120,0.011979555918110741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,5120,0.021056888831986323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,8192,0.02102222210831112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,4096,0.010272888673676385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,4096,0.018188445104493033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,3584,0.009553777674833933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,3584,0.016903999778959487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,3072,0.00941422250535753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,3072,0.008746667040718926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,3072,0.01551644504070282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,2560,0.008270222279760573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,2560,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,2560,0.013400889105266996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,512,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,2048,0.007859555383523306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,2048,0.011424889167149862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,1536,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,1536,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,1536,0.009905777871608734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,1024,0.005842666659090254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,1024,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,1024,0.008613333106040955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,256,0.005050666630268097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,768,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,768,0.0041724443435668945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,768,0.008056888977686564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,512,0.0037911112109820047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,512,0.005209777918126848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,512,0.007382222347789341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,256,0.0033946666452619764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,256,0.006810666786299811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,128,0.004552888787455029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,128,0.003113777687152227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,128,0.00639111093348927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1536,128,0.004770666774776247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,64,0.0029608888758553397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,64,0.006498666687144174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,1024,32,0.0030906666070222855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,1024,32,0.006478222293986215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,65536,0.06595200300216675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,65536,0.10490577750735813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,16384,0.0299617780579461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,16384,0.025096888343493145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,65536,0.1918942266040378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,16384,0.05041689011785719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,12288,0.023434667123688593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,12288,0.021203555994563635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,12288,0.03970044520166185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,10240,0.020655999581019085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,10240,0.01993777851263682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,10240,0.03458222084575229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,8192,0.01647200021478865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,8192,0.028970665401882593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,7168,0.015976889265908133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,7168,0.015070221490330167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,7168,0.026348445150587294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,6144,0.014404444230927361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,6144,0.013478221992651621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,6144,0.02365066607793172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,5120,0.012974222501118978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,5120,0.01183555523554484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,5120,0.021126222279336717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,4096,0.01110311100880305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,4096,0.010096000300513374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,4096,0.01817599932352702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,16384,0.03044711218939887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,3584,0.009349333743254343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,3584,0.01681866745154063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,12288,0.023410666320059035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,3072,0.00853777759604984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,3072,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,4096,0.011391110718250275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,2560,0.008290666672918532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,2560,0.013225778109497495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,2048,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,2048,0.007144889069928064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,2048,0.011356444822417365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,1536,0.006485333459244833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,1536,0.005719999886221356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,1536,0.009878222313192155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,3584,0.010686222049925061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,1024,0.004641777939266629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,1024,0.008575999902354347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,768,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,768,0.004206222378545337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,768,0.008029333419269985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,2048,0.007344889144102733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,512,0.003752888904677497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,512,0.007345777418878343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,256,0.004724444614516364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,256,0.0032924444725116095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,256,0.006754666566848755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,1024,256,0.0047848886913723415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,8192,0.017276444368892245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,128,0.003040888864133093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,128,0.00637511122557852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,64,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,64,0.006487111250559489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,768,32,0.002936888900068071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,768,32,0.006445333361625671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,3584,0.009946666657924652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,3072,0.008816888762844933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,65536,0.051339556773503624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,16384,0.027631110615200464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,16384,0.01826133330663045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,16384,0.05068622363938225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,12288,0.02169955604606205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,65536,0.18893955813513863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,12288,0.016549333930015564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,12288,0.04031733340687222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,10240,0.014880888991885714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,10240,0.018941332896550495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,10240,0.03491377830505371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,8192,0.016198222835858662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,8192,0.013120888835854001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,8192,0.029215110672844782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,7168,0.014902222487661572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,7168,0.026313778426912095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,7168,0.01223288890388277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,6144,0.013594666288958656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,6144,0.023771555887328252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,6144,0.012112889024946423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,5120,0.010843555960390302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,2560,0.00794222205877304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,5120,0.021008888880411785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,4096,0.010228444304731157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,4096,0.009263111485375298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,4096,0.017973333597183228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,3584,0.009163555171754625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,3584,0.008596444295512306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,3584,0.015989333391189575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,3072,0.008183110919263627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,3072,0.007857778006129796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,3072,0.013936888840463428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,2560,0.00756444450881746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,2560,0.0075866662793689305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,2560,0.012368000215954252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,2048,0.00685777763525645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,2048,0.006511999915043513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,2048,0.010977778169843884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,1536,0.005498666730191972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,1536,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,1536,0.009713778065310584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,1024,0.004598222259018156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,1024,0.00849422232972251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,768,0.005296000176005893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,768,0.0041777777175108595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,768,0.007954667011896769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,512,0.003671110918124517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,512,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,256,0.00452533322903845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,256,0.0032142222755485107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,256,0.006699555450015598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,128,0.004319111092223061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,128,0.002934222213096089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,128,0.006344888773229387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,64,0.0027439999911520216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,64,0.006411555740568373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,512,32,0.002785777880085839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,512,32,0.0064204442832205035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,65536,0.09132977988984849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,65536,0.036694222026401095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,65536,0.1910782257715861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,16384,0.025967111190160114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,16384,0.013263999587959714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,16384,0.05055110984378391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,12288,0.02051555613676707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,12288,0.012988444831636218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,12288,0.0395760006374783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,10240,0.018020444446139865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,10240,0.011838222543398539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,10240,0.03433422247568766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,8192,0.015250666273964776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,8192,0.010598222414652506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,8192,0.02885155545340644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,1024,0.0057795556883017225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,7168,0.010328888893127441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,7168,0.026293334033754136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,6144,0.01277599980433782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,6144,0.009496000077989366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,6144,0.02370044423474206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,512,0.005117333183685939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,5120,0.00980266680320104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,5120,0.02092711130777995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,4096,0.009383999639087254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,4096,0.00869777798652649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,4096,0.01703377730316586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,768,128,0.004456888884305954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,3584,0.008194666769769456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,3584,0.015415999624464246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,65536,0.09660889042748345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,3072,0.006893333461549547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,3072,0.013608888619475894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,5120,0.012125333150227865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,1024,0.005541333307822545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,2560,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,2560,0.01219466659757826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,512,512,0.004909333255555895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,2048,0.005495111147562663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,2048,0.010942222343550788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,1536,0.004772444566090901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,1536,0.009736888938479954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,1024,0.005428444594144821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,1024,0.004096888833575779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,1024,0.008466666771305932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,7168,0.014180445008807711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,768,0.0037964445849259696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,768,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,512,0.004830222162935468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,512,0.003458666718668408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,512,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,256,0.004534222185611725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,256,0.0031022222505675424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,256,0.006690666493442323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,128,0.004402666870090696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,128,0.0029155556112527847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,128,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,64,0.0027297778675953546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,64,0.0064222220745351576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,256,32,0.002752888947725296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,256,32,0.006406222366624408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,65536,0.09220088852776422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,65536,0.03226222263442145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,65536,0.1942462258868747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,16384,0.02606755495071411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,16384,0.013571555415789286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,16384,0.05114577876196968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,12288,0.020672889219390023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,12288,0.011499555574523078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,5120,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,12288,0.0395431121190389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,10240,0.018039999736679923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,10240,0.010632000035709806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,10240,0.03446844551298354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,8192,0.015139554937680563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,8192,0.009665778113736046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,8192,0.028733332951863606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,7168,0.013984888792037964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,7168,0.009158222211731805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,7168,0.026194666822751362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,6144,0.012147555748621622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,6144,0.008349333372380998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,6144,0.023524444964196947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,5120,0.007243555453088548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,5120,0.02030933399995168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,3584,0.008743111458089616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,4096,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,4096,0.016343111793200176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,3072,0.007884444461928474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,3584,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,3584,0.015141333142916361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,3072,0.007735110819339752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,3072,0.006663110935025745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,3072,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,2560,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,2560,0.00619911112719112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,2560,0.012267555627557965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,2048,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,2048,0.005449777675999536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,2048,0.01090222183201048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,1536,0.006078222145636876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,1536,0.004777777940034866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,1536,0.009721777505344814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,1024,0.0053635554181204895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,1024,0.0040755557517210645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,1024,0.00848444468445248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,768,0.005071111023426056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,768,0.003742222156789568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,768,0.007872888611422645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,512,0.0047493332790003884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,512,0.0034479999707804788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,512,0.007225777539942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,256,768,0.005136000199450387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,256,0.003061333257291052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,256,0.006671111202902264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,128,0.0041848888827694785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,128,0.0028604444944196274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,128,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,64,0.0027146666414207886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,64,0.006364444477690591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,128,32,0.00272088891102208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,128,32,0.0063528890411059065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,65536,0.029115554359224107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,65536,0.1931555536058214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,16384,0.011560888753996955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,16384,0.05058133271005419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,12288,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,12288,0.039817776944902204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,10240,0.009436444275909001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,10240,0.03444177905718485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,8192,0.008544888761308458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,8192,0.028862221373452082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,7168,0.00816177742348777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,7168,0.026399999856948853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,6144,0.00720355576939053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,6144,0.023531556129455566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,5120,0.00703999979628457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,5120,0.02017422186003791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,4096,0.006520888871616787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,4096,0.016283555163277518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,3584,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,3584,0.014919110470347934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,3072,0.006611555400821898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,3072,0.013551111022631327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,2560,0.006055111272467508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,2560,0.012305777933862476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,2048,0.0053582220441765254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,2048,0.010996444357766045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,1536,0.004688888788223267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,1536,0.00982844415638182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,1024,0.003995555556482739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,5120,0.010793777803579966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,1024,0.008481777376598781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,768,0.003663111064169142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,768,0.007846222155623967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,512,0.003318222239613533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,256,0.002999111182159848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,512,0.007186666958861881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,256,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,128,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,128,0.002769777758253945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,64,0.0026186667382717133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,64,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,64,32,0.002647111192345619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,4096,0.009132444030708736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,64,32,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,65536,0.026068444053332012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,16384,0.010639111200968424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,12288,0.009218666288587783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,16384,0.051029334465662636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,10240,0.008828444613350762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,12288,0.039777778916888766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,8192,0.007993777592976889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,10240,0.03418755531311035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,65536,0.19328533278571236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,7168,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,8192,0.028699556986490887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,6144,0.006663110935025745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,6144,0.02334044377009074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,7168,0.026166222161716882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,5120,0.006671999891599019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,4096,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,5120,0.019699555304315355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,4096,0.015861334072219003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,3584,0.006820444431569841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,3072,0.006375999914275275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,3584,0.014767110347747803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,3072,0.013444444371594323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,2560,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,2560,0.012164444559150271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,2048,0.010856000085671743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,2048,0.005364444520738389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,1536,0.00462755560874939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,1536,0.009658666948477427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,1024,0.004007110993067424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,1024,0.008382221890820397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,768,0.0035982223020659555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,768,0.007876444194051955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,512,0.003286222202910317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,512,0.007106666763623555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,256,0.0029964444951878656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,128,0.00272444449365139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,256,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,128,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,64,0.0025946667624844448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,64,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,384,32,32,0.0026444445053736367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,384,32,32,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,3584,0.00852711084816191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,16384,0.7941910955641004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,2560,0.007139555282062954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,16384,0.4477795494927301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,12288,0.6199813418918186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,16384,0.8684506416320801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,10240,0.5319235589769151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,12288,0.6924764315287272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,10240,0.28756533728705513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,8192,0.4077226585812039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,8192,0.25139376852247447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,10240,0.55931642320421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,8192,0.4456479814317491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,7168,0.3589955435858832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,7168,0.20604356129964194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,6144,0.31071289380391437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,7168,0.38449422518412274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,6144,0.1753680043750339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,5120,0.26263199912177193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,2048,0.00654755574133661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,5120,0.14950666162702772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,6144,0.3164409001668294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,1536,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,4096,0.21411555343204072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,3584,0.19365510675642225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,5120,0.25953600141737193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,3584,0.11051644219292535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,4096,0.21403643820020887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,3072,0.16715910699632433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,3072,0.09646222326490615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,3584,0.1785573297076755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,384,128,256,0.004383111165629493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,3072,0.14384710788726807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,2560,0.14237688647376165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,2048,0.11650666925642227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,2560,0.1115671131345961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,2048,0.06786666976081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,1536,0.09249066644244724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,2048,0.09822755389743382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,1536,0.054915553993648954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,1536,0.06623644298977323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,1024,0.06517689095603095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,768,0.05299466517236498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,768,0.0358168880144755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,1024,0.055252446068657764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,512,0.040511998865339495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,512,0.029709332519107397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,768,0.04764266808827718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,512,0.04286311070124308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,256,0.029495110114415485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,128,0.018236445056067575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,256,0.04107022285461426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,128,0.021889777647124395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,64,0.016325333052211337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,128,0.04135288794835409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,65536,32,0.017047999633683097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,64,0.04308088951640659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,65536,32,0.04642666710747612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,65536,0.8343333138359917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,65536,0.45820977952745223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,12288,0.32456888092888725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,16384,0.21301689412858751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,16384,0.18680532773335776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,12288,0.09389688571294148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,12288,0.1549866729312473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,12288,0.15101422203911677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,10240,0.07643022139867146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,10240,0.1342702176835802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,65536,0.8967866897583008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,10240,0.1346906688478258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,8192,0.0627760026190016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,8192,0.10893244213528103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,8192,0.10329688919915093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,7168,0.05647999710506863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,7168,0.0967857771449619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,6144,0.08414222134484185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,6144,0.04942933387226529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,7168,0.08039377795325385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,5120,0.0681848857137892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,5120,0.04283466604020861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,6144,0.06943821907043457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,4096,0.03589066531923082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,5120,0.05499644411934746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,4096,0.06105689207712809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,3584,0.05011466807789273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,3584,0.032863110303878784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,4096,0.04416711131731669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,3584,0.03876533442073398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,3072,0.04422577884462145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,3072,0.03453155689769321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,2560,0.026591999663247004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,2560,0.03788977861404419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,2560,0.030643555853101943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,2048,0.02244088881545597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,2048,0.03151999910672506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,2048,0.026175111532211304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,1536,0.019222221440739103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,4096,0.12141156196594238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,1536,0.025618667403856914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,1536,0.02136444383197361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,1024,0.01792977750301361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,1024,0.01608000033431583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,1024,0.01759111053413815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,768,0.014326221413082547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,768,0.01460088955031501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,512,0.01108088923825158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,768,0.016016888949606154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,512,0.014740443891949125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,256,0.010976889067225985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,256,0.008441777692900764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,256,0.012816000315878125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,128,0.0064693333374129395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,128,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,64,0.005382222019963794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,64,0.011734222372372946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,16384,32,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,16384,32,0.014011555247836642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,2560,0.08307644393708971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,1024,0.04355288876427544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,65536,256,0.02609955602222019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,65536,0.6050364706251357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,65536,0.33829598956637913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,16384,0.08906222052044338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,16384,0.16783377859327528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,12288,0.06896622313393487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,12288,0.1180924442079332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,16384,0.1373128890991211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,12288,0.11511645052168105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,10240,0.09698133336173163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,65536,0.7344417572021484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,10240,0.10211377673678929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,8192,0.07866666714350383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,8192,0.07542666461732653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,7168,0.04420800010363261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,16384,0.11964800622728135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,7168,0.0694284439086914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,7168,0.06367822488149007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,6144,0.03896533449490865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,6144,0.06153066953023275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,6144,0.05338933401637607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,5120,0.051028443707360156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,5120,0.0425546665986379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,4096,0.028456889920764502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,4096,0.04196977615356445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,4096,0.03441155619091458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,3584,0.03800444470511542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,3584,0.030559112628300984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,3072,0.023730667101012334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,3072,0.03372977839575873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,3072,0.02764355474048191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,2560,0.02093422247303857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,2560,0.028839998775058325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,2560,0.024460444847742718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,2048,0.018043556147151522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,2048,0.024104000793562994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,2048,0.021306667062971327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,3072,0.02980355421702067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,1536,0.01941866676012675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,1536,0.016217778126398723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,1536,0.017614222235149808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,1024,0.013350221845838757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,1024,0.01348355578051673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,1024,0.014732443624072604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,768,0.011323555476135679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,768,0.013590222431553734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,512,0.010944888823562197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,512,0.009159999589125315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,512,0.012703110774358114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,512,0.012668444050682915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,256,0.007116444408893585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,256,0.009844444692134857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,256,0.009869332942697737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,128,0.008454221818182204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,128,0.005164444446563721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,64,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,128,0.00942311104800966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,12288,32,0.005007110950019625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,64,0.009639110830095079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,12288,32,0.011568889021873474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,16384,128,0.00959911114639706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,65536,0.532507578531901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,16384,0.0936959981918335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,16384,0.1409671174155341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,65536,0.6067048708597819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,16384,0.11730755700005426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,12288,0.06765600045522054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,12288,0.09956266482671101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,12288,0.09471911191940308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,10240,0.052242666482925415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,10240,0.08261777957280476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,8192,0.043056888712777026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,10240,0.08474489053090413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,10240,0.058761775493621826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,8192,0.06724622514512804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,7168,0.03931377662552728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,7168,0.05957688887914022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,8192,0.0565813316239251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,6144,0.05231111248334249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,7168,0.04971288972430759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,6144,0.03451022174623277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,5120,0.044587555858823985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,6144,0.044778668218188815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,8192,0.049039999643961586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,4096,0.025268443756633337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,5120,0.03576088945070902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,4096,0.036850667662090726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,3584,0.033306668202082314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,3584,0.02346844474474589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,4096,0.029734220769670274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,3584,0.02700977855258518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,3072,0.029655999607510034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,3072,0.020745777421527438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,2560,0.0255404445860121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,3072,0.023942222197850544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,2560,0.018527110417683918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,2048,0.021330666210916307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,2560,0.021727111604478624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,2048,0.01645155582163069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,2048,0.018975110517607797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,1536,0.01682844426896837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,1536,0.015585778488053216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,1024,0.01200799975130293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,1024,0.01349955548842748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,1024,0.012368888490729861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,768,0.010802666346232096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,768,0.011217777927716574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,768,0.012520888613329994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,512,0.009879111415810054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,512,0.008951110972298516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,512,0.01106844428512785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,256,0.008777778181764815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,256,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,256,0.009198221895429824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,128,0.007448888487286038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,128,0.005483555710977978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,5120,0.03379377722740173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,128,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,64,0.00906044410334693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,32,0.00490399988161193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,10240,64,0.004613333278232151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,10240,32,0.01071911139620675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,65536,0.40873334142896867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,3584,0.026370666093296472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,16384,0.06406399938795301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,16384,0.10290222035513984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,16384,0.0946248902214898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,12288,0.04984800020853678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,65536,0.4647670851813422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,12288,0.07900000280804105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,10240,0.04275911053021749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,12288,0.07295555538601346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,10240,0.06565066840913561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,10240,0.06517955329683092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,8192,0.03552622265285916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,8192,0.05318311187956068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,8192,0.04794489012824165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,7168,0.047693332036336265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,7168,0.040207998620139226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,6144,0.028578665521409776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,12288,768,0.012264000044928657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,6144,0.04197333256403605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,6144,0.03680888811747233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,5120,0.035912000470691256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,5120,0.03009066647953457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,4096,0.020977778567208186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,4096,0.0296106669637892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,4096,0.025024889243973628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,3584,0.026816000541051228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,3584,0.02290133304066128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,3072,0.017720000611411203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,3072,0.023767110374238756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,3072,0.0203804439968533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,2560,0.016338666280110676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,2560,0.02051466703414917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,2560,0.018379555808173288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,2048,0.014389332797792224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,2048,0.01666755477587382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,2048,0.01612977849112617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,1536,0.012720000412729053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,1536,0.013212444053755866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,1536,0.01370933320787218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,1024,0.010815111299355825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,1024,0.009929777847396003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,1024,0.01200799975130293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,768,0.010097777677906884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,768,0.008744888835483128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,768,0.011267555256684622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,512,0.009182222187519073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,512,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,512,0.009304888546466827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,256,0.007888889147175683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,256,0.005136888888147143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,256,0.00830577810605367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,65536,0.28286578920152455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,128,0.0064995553758409284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,128,0.004334222111437055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,128,0.007882666256692674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,64,0.003936888857020272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,64,0.008328888979223039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,8192,32,0.004352000024583605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,8192,32,0.009122666385438707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,65536,0.3542293442620172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,65536,0.2324791219499376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,5120,0.030096888542175293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,16384,0.09154755539364284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,65536,0.4188008838229709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,12288,0.046557333734300405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,16384,0.08745866351657444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,12288,0.0695377786954244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,12288,0.07348889112472534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,10240,1536,0.014512888259357877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,10240,0.03993155558904012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,10240,0.059347556696997754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,8192,0.0329733325375451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,10240,0.06051466862360636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,8192,0.04724444283379448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,7168,0.04217244519127739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,7168,0.029785778787400987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,8192,0.04338933361901177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,6144,0.037794666157828435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,7168,0.03977155685424805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,6144,0.026149332523345947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,6144,0.034099555677837796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,5120,0.022834666901164587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,5120,0.03279555506176419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,5120,0.029169778029123943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,4096,0.01961599952644772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,4096,0.026662222213215295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,4096,0.024513777759340074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,3584,0.018407111366589863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,3584,0.02397866712676154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,3584,0.022800000177489385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,3072,0.016924444172117446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,3072,0.021680000755521987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,3072,0.019873778025309246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,2560,0.01850133306450314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,2560,0.015440000428093804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,2560,0.018010665973027546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,2048,0.013542222479979197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,2048,0.015073777900801765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,2048,0.0158906661801868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,1536,0.012279999752839407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,1536,0.012103111379676394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,1536,0.013803555733627744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,1024,0.009275555610656738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,1024,0.012058667010731168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,768,0.009692444569534725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,768,0.008026666939258575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,768,0.01093066649304496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,512,0.009129777550697327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,512,0.00674933319290479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,512,0.009014222357008193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,65536,0.4379724396599664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,256,0.01901244454913669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,256,0.013816888961527081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,128,0.006088888893524806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,128,0.004841777599520153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,128,0.009703999592198266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,64,0.004368888835112254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,64,0.008799110849698385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,7168,32,0.004302222281694412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,7168,32,0.009713778065310584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,7168,0.03235200047492981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,65536,0.3146897686852349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,5120,0.024710221423043147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,16384,0.07906577984491985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,8192,3584,0.01943111088540819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,16384,0.06752177741792467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,12288,0.04679022232691447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,65536,0.34909600681728786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,12288,0.06158488988876343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,12288,0.05520622266663445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,10240,0.055941330062018506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,10240,0.04025955663786994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,10240,0.04813244607713488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,8192,0.041887111133999295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,8192,0.03523644473817613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,7168,0.029912001556820337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,7168,0.03770222266515096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,7168,0.03190133306715224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,6144,0.026155556241671245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,6144,0.03369066781467862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,6144,0.02795999911096361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,5120,0.022856000396940444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,5120,0.02978666623433431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,5120,0.02443644404411316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,4096,0.01944977707333035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,4096,0.023617777559492324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,4096,0.020434666011068556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,3584,0.02121777832508087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,3584,0.018389332625601027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,3584,0.01871200071440803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,3072,0.01956888867749108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,3072,0.016897777716318767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,2560,0.015113777584499784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,2560,0.01625955601533254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,2560,0.015336889359686108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,2048,0.013316444224781461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,2048,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,2048,0.013791999883121915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,16384,0.06016266345977783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,1536,0.011054221954610614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,1536,0.011936888926559024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,1024,0.010886222124099731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,1024,0.008575110799736446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,7168,256,0.007146666447321574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,1024,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,768,0.0074177781740824384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,768,0.009371555513805812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,768,0.009510222408506606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,512,0.008021333151393467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,512,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,512,0.008151111503442129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,256,0.004567111117972268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,256,0.007500444849332173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,128,0.005480000128348668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,128,0.004060444318585926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,128,0.007165333463086023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,64,0.00369777778784434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,64,0.007313778003056844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,6144,32,0.003954666770166821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,6144,32,0.008078222473462423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,65536,0.19127021895514595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,65536,0.27905599276224774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,65536,0.47484800550672746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,65536,0.3392746713426378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,16384,0.058543999989827476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,16384,0.054246223635143705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,16384,0.06897422340181139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,16384,0.0632408857345581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,12288,0.042452444632848106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,12288,0.056629333231184215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,12288,0.05146577623155382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,10240,0.03669155637423197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,10240,0.05115288827154371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,10240,0.046237332953347095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,8192,0.03311199943224589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,8192,0.03674222363366021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,8192,0.0344177782535553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,7168,0.027237332529491846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,7168,0.034953776333067156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,7168,0.030551112360424463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,6144,0.030830221043692693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,6144,0.0236462222205268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,6144,0.02739288906256358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,5120,0.02087200019094679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,5120,0.02769511110252804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,5120,0.024012444747818842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,4096,0.0181057784292433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,4096,0.02130488885773553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,4096,0.01997244523631202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,3584,0.016720000240537856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,3584,0.019481778144836426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,3584,0.01865511139233907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,3072,0.015320888823933072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,3072,0.017595556047227647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,3072,0.01649777756796943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,3072,0.01661688917213016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,2560,0.014905777242448596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,2560,0.015007111761305066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,2048,0.012529777983824411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,2048,0.01275199982855055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,2048,0.014013333453072442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,1536,0.010949333508809408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,1536,0.010481778118345471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,1536,0.011847111086050669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,1024,0.009486222432719337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,1024,0.008187555604510838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,1024,0.009851555857393477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,768,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,768,0.007117333511511485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,768,0.008594666918118795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,512,0.007983110845088959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,512,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,256,0.00434755575325754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,256,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,256,0.007516444557242923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,128,0.00517866677708096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,128,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,1536,0.011951111257076263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,64,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,128,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,64,0.007394666473070781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,5120,32,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,5120,32,0.0038408889538711975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,65536,0.21548710929022896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,1024,0.010268444816271463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,65536,0.15167110496097141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,16384,0.042152000798119434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,16384,0.061773333284589976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,65536,0.23249422179328072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,12288,0.0532915559079912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,16384,0.046550220913357206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,12288,0.03157600098186069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,10240,0.04708266589376661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,12288,0.03743733300103082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,10240,0.0336444444126553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,8192,0.032877332634396024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,8192,0.023711999257405598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,8192,0.02549244463443756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,7168,0.031457778480317854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,7168,0.02137600051032172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,7168,0.023382221659024555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,6144,0.028239998552534316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,6144,0.01920977731545766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,6144,0.021006221572558086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,5120,0.016965332958433364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,5120,0.024970667229758367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,5120,0.018745778335465323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,4096,0.018882667024930317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,4096,0.0148026661740409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,4096,0.01607733302646213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,3584,0.01738133364253574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,3584,0.014719999498791166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,3072,0.012797333300113678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,3072,0.015639111399650574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,3072,0.013798221945762634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,2560,0.011688888900809817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,2560,0.013543999857372709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,2560,0.012532444463835822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,2048,0.010791111323568555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,2048,0.011271110839313932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,2048,0.011623111036088733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,1536,0.00960711141427358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,1536,0.009456000394291347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,1536,0.009936889012654623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,1024,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,1024,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,1024,0.007901333272457123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,768,0.006694222076071634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,768,0.006215999937719769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,768,0.0075137780772315125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,512,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,512,0.004895110925038655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,512,0.006816888848940532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,256,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,256,0.003990222182538775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,256,0.0064222220745351576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,128,0.0048142220411035745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,128,0.003569777641031477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,128,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,64,0.0033093332830402586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,64,0.006435555716355641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,4096,32,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,4096,32,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,6144,256,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,65536,0.18850310643513998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,65536,0.23536711268954805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,16384,0.041160888142055936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,16384,0.05556888712777031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,16384,0.04518933428658379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,12288,0.04564266734653049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,12288,0.03274222215016683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,12288,0.03654222355948554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,10240,0.02865866820017497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,10240,0.03622488843070136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,10240,0.03272177775700887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,8192,0.023610666394233704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,8192,0.029029332929187354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,8192,0.025352888637118872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,7168,0.02208711041344537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,7168,0.026703110999531213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,7168,0.023285332653257582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,6144,0.01905955539809333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,6144,0.023530667026837666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,6144,0.020625778370433383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,5120,0.02010222276051839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,5120,0.017133333616786532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,5120,0.018724444839689467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,4096,0.014896000425020853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,4096,0.017391110459963482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,4096,0.015824000040690105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,3584,0.014173333843549093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,3584,0.014944000376595391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,3584,0.014946666028764514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,3072,0.012862222062216865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,3072,0.013595555391576556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,3072,0.013552889227867126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,2560,0.012469333079126147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,2560,0.01256444470749961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,2048,0.010799999866220685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,2048,0.010570666856235929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,2048,0.01164177805185318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,1536,0.009778666827413771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,1536,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,1536,0.010012444522645738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,1024,0.008382221890820397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,1024,0.00702400008837382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,1024,0.00797155582242542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,768,0.006913777854707506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,768,0.006029333505365584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,768,0.007391999993059371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,8192,0.030176000462638006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,512,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,512,0.0047120000753137796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,512,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,256,0.0037804444630940757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,256,0.0053680001033677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,256,0.006532444308201472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,128,0.005027555757098728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,128,0.0033928888539473214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,128,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,64,0.0031902222997612427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,64,0.006525333142942852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3584,32,0.003383111208677292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3584,32,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,65536,0.1637680000729031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,2560,0.01387733303838306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,65536,0.2286133236355252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,16384,0.04933422141604953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,16384,0.044554667340384595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,5120,512,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,12288,0.041731556256612144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,16384,0.04515377680460612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,12288,0.033751110235850014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,10240,0.03562755717171563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,10240,0.032031112247043185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,10240,0.027594667341974046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,8192,0.027992000182469685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,8192,0.024859555893474158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,4096,3584,0.013813333378897773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,7168,0.024925332930352952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,7168,0.022632888621754114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,6144,0.01962311069170634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,6144,0.02218311031659444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,65536,0.15261599752638075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,6144,0.02123111155298021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,5120,0.017403556240929496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,5120,0.01938666734430525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,5120,0.01809866726398468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,4096,0.015055111712879606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,4096,0.015815110670195687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,4096,0.015781333049138386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,3584,0.013962666193644205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,3584,0.013960889644092984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,3584,0.014461333552996317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,3072,0.012723555167516073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,3072,0.012658666405412885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,3072,0.013384888569513956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,2560,0.011732444167137146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,2560,0.010625777973069085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,2560,0.012350222302807702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,2048,0.010300444232092964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,2048,0.00923822240697013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,2048,0.011185777684052786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,1536,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,1536,0.007832888927724626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,1536,0.009678222239017487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,1024,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,1024,0.006360000206364526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,1024,0.007851555943489075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,768,0.0058346668051348785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,768,0.005139555368158552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,768,0.007352888584136963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,512,0.0041884444653987885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,512,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,512,0.006864000111818314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,256,0.004959110998445087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,256,0.0036355555057525635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,256,0.006516444600290722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,128,0.0046568889584806235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,128,0.0032995556377702286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,128,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,64,0.003107555624511507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,64,0.006467555546098285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,3072,32,0.0033928888539473214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,3072,32,0.006832888970772426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,65536,0.14040977425045437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,65536,0.12443110677931045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3584,2560,0.011983999775515662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,16384,0.03454044461250305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,16384,0.043102221356497876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,16384,0.04302577839957344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,65536,0.20025244024064806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,12288,0.02699022160636054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,12288,0.035968889792760216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,10240,0.03257511059443156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,10240,0.02352711061636607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,12288,0.03367200162675645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,10240,0.02942755487230089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,8192,0.02310488952530755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,8192,0.02008888953261905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,8192,0.02464444438616435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,7168,0.020798222886191476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,7168,0.022664000590642292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,6144,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,6144,0.018132444885041978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,6144,0.02016888890001509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,5120,0.014521777629852295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,5120,0.015403555499182807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,5120,0.018226666582955253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,4096,0.0138568886452251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,4096,0.015771556231710646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,65536,0.15724710623423258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,3584,0.012637333737479316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,3584,0.014481777946154276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,3072,0.011138666835096149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,3072,0.011799111134476133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,3072,0.013297778036859302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,12288,0.03438577718204922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,2560,0.010038221875826517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,2560,0.012402666939629449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,10240,0.029326223664813574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,2048,0.008619555168681675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,2048,0.011128889189826118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,1536,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,1536,0.007428444094128079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,8192,0.024167999625205994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,1536,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,1024,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,1024,0.007920000288221572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,768,0.006124444305896759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,768,0.004631999880075455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,768,0.007443555527263218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,512,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,512,0.004060444318585926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,512,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,3072,7168,0.022208889325459797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,256,0.0035377778112888336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,256,0.0064497776329517365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,4096,0.01277955538696713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,7168,0.018277333842383493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,128,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,128,0.0031848889258172777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,64,0.0064444442590077715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,64,0.003064888839920362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2560,32,0.003121777748068174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2560,32,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,3584,0.012148444851239523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,65536,0.10341066784328884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,65536,0.11371910572052002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,65536,0.12056533495585124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,16384,0.03660800059636434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,16384,0.028299556838141546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,2560,0.010431999961535135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,2048,0.009400000174840292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,12288,0.022843556271659002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,12288,0.031065778599845037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,10240,0.029409776131312054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,10240,0.020215999748971727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,8192,0.017447110679414537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,1024,0.006707555717892117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,8192,0.020432889461517334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,8192,0.01736088924937778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,7168,0.020757334099875558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,7168,0.015976000163290236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,7168,0.016171556380059984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,6144,0.014217777384652032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,6144,0.018037334084510803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,6144,0.014856888188256158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,5120,0.015746666325463187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,5120,0.013605333036846586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,256,0.005124444348944558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,4096,0.011454222102959951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,4096,0.012091555529170565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,4096,0.011878222227096558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,3584,0.010775110787815519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,3584,0.011326221956147088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,3072,0.010057777994208867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,3072,0.010016000105275048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2560,128,0.004774222357405556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,3072,0.010513777534166971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,2560,0.009192888935407003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,2560,0.009100444614887238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,2560,0.009735110733244155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,2048,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,2048,0.008188444707128737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,1536,0.006504000061088138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,1536,0.007083555890454187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,1024,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,1024,0.00498933345079422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,1024,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,16384,0.029294222593307495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,768,0.004376000000370873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,768,0.006416888700591193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,512,0.005120888766315249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,512,0.003850666599141227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,512,0.006068444500366847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,256,0.004755555755562252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,256,0.0034462221794658234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,256,0.005774222314357758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,12288,0.023240000009536743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,128,0.0032088889016045462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,128,0.005647111270162795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,64,0.003083555648724238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,64,0.005850666513045629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,2048,32,0.0032622222271230486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,2048,32,0.006036444256703059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,10240,0.020421332783169217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,65536,0.08980977535247803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,65536,0.09736266401078965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,5120,0.012821333275900947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,16384,0.030236442883809406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,16384,0.02661777701642778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,3584,0.010936888555685679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,12288,0.02671555512481266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,12288,0.021980444590250652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,1536,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,10240,0.023846222294701472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,2048,0.008103110724025303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,10240,0.01941511034965515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,8192,0.01660977800687154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,8192,0.01702666613790724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,768,0.005518222020732031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,7168,0.015307555596033732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,8192,0.016545777519543965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,7168,0.015623999966515435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,7168,0.015683554940753512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,6144,0.013755555782053204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,6144,0.014309333430396186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,5120,0.012330667012267642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,5120,0.014783110883500842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,5120,0.013236444857385425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,4096,0.011015111373530494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,4096,0.011682666838169098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,4096,0.01164533363448249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,3584,0.01073688848151101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,3584,0.010484444598356882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,3584,0.011066666907734342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,3072,0.00962577760219574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,3072,0.009679999616410997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,3072,0.010409778190983666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,2560,0.008434666527642144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,2560,0.008960888617568547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,2560,0.008219555848174626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,2048,0.007273777491516537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,2048,0.007730666961934831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,1536,0.00646666685740153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,1536,0.006045333213276333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,1536,0.0069982219073507525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,1024,0.005856888989607493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,1024,0.004962666581074397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,1024,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,2048,128,0.004509333521127701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,768,0.00443733317984475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,768,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,512,0.005081777771313985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,512,0.003942222230964237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,512,0.006032889087994893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,256,0.004775111046102312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,256,0.0032835555159383347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,256,0.005728888842794631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,128,0.004554666578769684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,128,0.003023111157947116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,128,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,64,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,64,0.005699555493063397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1536,32,0.0030622221529483795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1536,32,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,65536,0.10245955652660793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,65536,0.06572800212436251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,65536,0.08586577574412028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,16384,0.02812355425622728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,16384,0.026076444321208533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,16384,0.02530133393075731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,16384,0.02500088844034407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,12288,0.02032088902261522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,12288,0.021736888421906367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,12288,0.021977777282396953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,12288,0.02035199933581882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,10240,0.018008889423476327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,10240,0.01793866687350803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,10240,0.018393778138690524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,8192,0.015631111131774057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,8192,0.015388444066047668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,8192,0.01577333278126187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,7168,0.014021333720948962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,7168,0.01442311041884952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,7168,0.014751111467679342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,6144,0.014640889234013028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,6144,0.013604444762070974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,5120,0.0129457778400845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,5120,0.011607999602953592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,5120,0.012792888614866467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,4096,0.009944889280531142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,4096,0.010283555421564314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,4096,0.011173333558771344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,3584,0.009663999908500248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,3584,0.009881777895821465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,3072,0.008169777691364288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,10240,0.019408888287014432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,3072,0.008859555754396651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,3072,0.009232000344329411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,2560,0.007385777930418651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,2560,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,2560,0.008177777959240807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,2048,0.006706666615274217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,2048,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,2048,0.007428444094128079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,1536,0.0061048890153567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,1536,0.0058017778727743365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,1536,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,1024,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,1024,0.006432000133726332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,6144,0.013599110974205865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,768,0.004143110993835661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,768,0.006217777729034424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,512,0.004910222358173794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,512,0.003664888855483797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,512,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,256,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,2048,0.007378666765160031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,256,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,256,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,128,0.004360888981156879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,128,0.005500444521506627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,128,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,64,0.0027911110470692315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,64,0.005547555370463266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,1024,32,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,1024,32,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,65536,0.05359911256366306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,65536,0.08667555782530044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,16384,0.025436444414986506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,65536,0.08752533462312485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,16384,0.018578666779730055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,16384,0.02518133322397868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,12288,0.015012444721327888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,12288,0.020449777444203694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1536,768,0.005494222044944763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,10240,0.014953777194023132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,10240,0.01815111107296414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,8192,0.01538044379817115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,8192,0.013167999684810638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,8192,0.015607999430762397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,7168,0.014259555273585849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,7168,0.01236977759334776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,7168,0.014741332994567024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,65536,0.08753422233793472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,6144,0.01222400036123064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,6144,0.0136808885468377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,5120,0.011240888800885944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,5120,0.010791111323568555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,5120,0.012676444318559436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,4096,0.009196444518036312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,4096,0.009343999955389235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,4096,0.010693333215183683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,6144,0.01276000009642707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,3584,0.00868800034125646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,3584,0.009664889011118147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,3072,0.007872888611422645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,3072,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,3072,0.008817777865462834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,2560,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,2560,0.007245333658324347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,3584,0.009183111290136972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,2560,0.007901333272457123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,2048,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,2048,0.007358222372002072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,1536,0.006080889039569431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,1536,0.005174222091833751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,1536,0.0069439998931354946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,1024,0.00564444437623024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,1024,0.004398222184843487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,1024,0.006487111250559489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,1024,768,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,768,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,768,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,768,0.004008000095685323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,512,0.004922666483455234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,512,0.0035946667194366455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,512,0.005988444305128521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,256,0.0032595555401510666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,256,0.004501333253251182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,256,0.005723555468850666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,128,0.004330666528807746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,128,0.0029964444951878656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,128,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,64,0.0028719999310043124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,768,32,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,64,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,768,32,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,65536,0.04964800013436211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,65536,0.08186488681369357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,10240,0.01772622267405192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,16384,0.0148062225845125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,12288,0.020074667202101815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,16384,0.02484888831774394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,12288,0.01904444396495819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,12288,0.012900444368521372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,12288,0.019832000136375427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,10240,0.016882666283183627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,10240,0.01164533363448249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,10240,0.017803554733594257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,8192,0.01019377758105596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,8192,0.015506666567590503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,7168,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,7168,0.010414222048388587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,7168,0.014473777678277759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,6144,0.012647999657524956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,6144,0.012425777812798819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,6144,0.013535999589496188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,3584,0.008683555656009251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,5120,0.009167110754383935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,5120,0.012130666938092975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,4096,0.008723555339707268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,4096,0.008981333010726506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,4096,0.010118222071064843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,3584,0.008108444511890411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,3584,0.008028444316652086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,3584,0.008949333594905006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,3072,0.007558222446176741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,3072,0.007506666911972894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,3072,0.00833333366447025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,2560,0.007035555938879649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,2560,0.006684444430801604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,2560,0.007786666353543599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,2048,0.006476444502671559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,2048,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,2048,0.007266666326257918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,1536,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,1536,0.004905777672926585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,1536,0.006780444333950679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,1024,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,1024,0.0041884444653987885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,1024,0.0063493334584765965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,768,0.005112888912359874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,1024,0.005547555370463266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,768,2048,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,768,0.003843555433882607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,768,0.006156444549560547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,512,0.0047528888616296975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,512,0.005853333406978183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,512,0.003506666670242945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,256,0.004495111190610462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,256,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,256,0.003145777723855443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,128,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,128,0.005470222069157495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,64,0.0027786667148272195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,64,0.005506666584147348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,512,32,0.0027724444452259275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,512,32,0.005507555686765247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,65536,0.08178577820460002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,65536,0.03676711188422309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,65536,0.08183911111619738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,16384,0.023792889383104112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,16384,0.012333333492279053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,16384,0.023891554938422307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,65536,0.08256000280380249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,12288,0.010852444503042432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,12288,0.019745777050654095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,16384,0.024209777514139812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,10240,0.010394666757848527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,10240,0.017753778232468497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,8192,0.014547554983033074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,8192,0.008745777938101027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,8192,0.015254222684436373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,7168,0.012705778082211813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,7168,0.008377778033415476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,7168,0.014256000518798828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,6144,0.011759111450778114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,6144,0.010040889183680216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,6144,0.013100444442696042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,5120,0.009107555780145857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,5120,0.01124888906876246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,4096,0.008526222573386298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,4096,0.007792889244026608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,4096,0.009204444785912832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,3584,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,3584,0.007337777978844113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,3584,0.008676444490750631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,3072,0.007385777930418651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,3072,0.00673333348499404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,3072,0.008231110870838165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,2560,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,2560,0.00609422226746877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,2560,0.007752888732486301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,2048,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,2048,0.00546577779783143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,2048,0.007205333146784041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,1536,0.0058115555180443665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,1536,0.004770666774776247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,1536,0.00675022229552269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,1024,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,1024,0.004081777814361784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,1024,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,768,0.004982222285535601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,768,0.0037368887828456033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,768,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,512,0.004670222186379963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,512,0.003375111147761345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,512,0.0058000000814596815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,256,0.004383111165629493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,256,0.003012444410059187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,256,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,128,0.004223111189073986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,128,0.002842666581273079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,128,0.005406222409672207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,64,0.0028400001012616684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,64,0.00544177782204416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,256,32,0.0029271110478374693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,256,32,0.005461333526505365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,65536,0.08181777927610609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,65536,0.02561155623859829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,65536,0.08368621932135688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,16384,0.023707555400000677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,16384,0.010782221953074137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,16384,0.0239991115199195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,12288,0.01871466636657715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,12288,0.00963466697269016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,12288,0.019701333509551156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,10240,0.016361777981122334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,10240,0.008984000318580205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,10240,0.017648888958825003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,8192,0.013630222115251752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,8192,0.007833778030342525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,8192,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,7168,0.012294222083356647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,7168,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,7168,0.013565333353148567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,6144,0.010928000013033548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,6144,0.006776888751321369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,6144,0.01178311142656538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,5120,0.009544000029563904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,5120,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,5120,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,4096,0.008422222402360704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,4096,0.006509333435032103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,4096,0.009187555147541894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,3584,0.007874666816658443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,3584,0.006913777854707506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,3584,0.008754666480753157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,3072,0.007263999846246507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,3072,0.006568888823191325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,5120,0.009963555468453301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,3072,0.008289777570300633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,2560,0.006823110911581252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,2560,0.006035555568006303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,2560,0.007800888684060838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,2048,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,2048,0.005388444496525659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,2048,0.007270221908887227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,1536,0.005743999862008625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,1536,0.0047004446387290955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,1536,0.006774222271309958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,1024,0.005166222237878376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,1024,0.004015111260943943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,1024,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,768,0.0036720000207424164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,768,0.006080889039569431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,512,0.004603555632962121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,512,0.0033404444240861465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,512,0.0057751110030545135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,256,0.004341333276695675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,256,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,256,0.005513777749405966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,128,0.004222222086456087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,128,0.002776888923512565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,128,0.005416000054942236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,64,0.0026462222966882917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,64,0.005435555759403441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,128,32,0.0026524443593290118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,128,32,0.005392888767851724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,65536,0.02142488956451416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,65536,0.08368621932135688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,16384,0.009341333475377824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,16384,0.02401688860522376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,12288,0.008458666503429413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,12288,0.01957244508796268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,10240,0.007903111477692923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,10240,0.017622222503026325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,8192,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,8192,0.014722666806644864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,7168,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,512,128,0.004302222281694412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,7168,0.012789333032237159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,6144,0.006430222342411677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,6144,0.01110755569405026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,5120,0.00676711110605134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,5120,0.010299555957317352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,4096,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,3584,0.006812444577614467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,4096,0.0092622223827574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,3584,0.008663111262851292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,3072,0.006514666808976068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,3072,0.008207110895050896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,2560,0.00600355573826366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,2560,0.007672888537247975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,2048,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,2048,0.007167110840479533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,1536,0.004643555730581284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,1536,0.006699555450015598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,1024,0.003992889076471329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,1024,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,768,0.003666666646798452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,768,0.006054222169849608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,512,0.0032986667421129015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,512,0.005768888940413793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,256,0.0029680000411139596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,128,0.0027466666781240036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,256,0.005504000104135937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,128,0.005409777578380372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,64,0.0026231110095977783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,64,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,64,32,0.002609777781698439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,64,32,0.005416000054942236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,65536,0.019741333193249173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,16384,0.008887111312813228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,65536,0.08435466554429795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,16384,0.023944000403086346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,12288,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,12288,0.019443555010689628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,10240,0.00740444411834081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,8192,0.006796444455782573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,10240,0.01740444368786282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,7168,0.0064533332155810455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,8192,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,6144,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,7168,0.012653333445390066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,5120,0.00666222224632899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,6144,0.011552000211344825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,4096,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,5120,0.01015733347998725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,4096,0.00906133320596483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,3584,0.006712889091836081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,3072,0.0063795554969045855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,3584,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,3072,0.008151999778217739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,2560,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,2560,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,2048,0.005327111317051782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,1536,0.0046364445653226645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,2048,0.007124444676770105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,1536,0.006693333387374878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,1024,0.00398577791121271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,1024,0.006188444379303191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,768,0.003657777690225177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,768,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,512,0.0032871110985676446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,512,0.0057297779454125305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,256,0.0031235555393828284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,256,0.005467555589146084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,128,0.0027662221756246355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,128,0.00536088893810908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,64,0.0026071110947264563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,64,0.005388444496525659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,256,32,32,0.0026177778426143858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,256,32,32,0.00535999983549118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,12288,0.018767999278174508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,10240,0.016700444949997794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,8192,0.014330666926172046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,16384,0.9602532916598849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,16384,0.4982302453782823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,6144,0.011271110839313932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,12288,0.9927440219455295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,256,5120,0.009752888646390703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,12288,0.3822088771396213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,10240,0.48086754480997723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,16384,0.6587288644578722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,8192,0.3822471035851373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,12288,0.5239884588453505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,8192,0.26803467008802623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,7168,0.3370542261335585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,10240,0.4852106836107042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,8192,0.3274533218807644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,7168,0.23983555369906953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,6144,0.2945386568705241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,6144,0.20651377571953666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,7168,0.2928328779008654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,5120,0.24736356735229492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,5120,0.1711991098192003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,4096,0.21096355385250518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,6144,0.2722960048251682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,4096,0.13542044162750244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,5120,0.17292533980475533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,4096,0.1567653285132514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,3584,0.18729421827528211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,3072,0.15752977795071071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,3072,0.1024506688117981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,3584,0.14808177947998047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,2560,0.13486488660176596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,2560,0.08738488621181911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,3072,0.1040782199965583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,2048,0.11143110858069526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,2560,0.09219733211729263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,2048,0.06998488638136122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,1536,0.08885866403579712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,1024,0.06390488810009427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,1024,0.04000622365209792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,1536,0.05496177739567227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,1024,0.044603556394577026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,768,0.05204088820351494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,512,0.027688889039887324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,768,0.04025422202216254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,512,0.04028622309366862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,512,0.03591555688116286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,256,0.021164443757798936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,256,0.024904888537194993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,256,0.032503998941845365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,128,0.018414222531848483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,128,0.030470222234725952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,64,0.014397333065668741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,64,0.031326221095191106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,65536,32,0.014706666270891825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,65536,32,0.03676799933115641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,256,128,768,0.004892444445027245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,65536,0.7192755805121528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,65536,0.5100240177578396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,16384,0.1345635520087348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,16384,0.2075288825564914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,16384,0.15109777450561523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,10240,0.32537955707973903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,12288,0.10337689187791611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,12288,0.14301866955227321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,65536,0.7094391187032064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,10240,0.12845333417256674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,12288,0.11680266592237686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,10240,0.10467021995120579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,8192,0.10285511281755234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,8192,0.07201244433720906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,7168,0.06116000148985121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,3584,0.1190462244881524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,7168,0.0860053300857544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,7168,0.059802664650811084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,6144,0.055323554409874805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,6144,0.07648799816767375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,6144,0.05766044722663032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,2048,0.07083377573225233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,5120,0.06417510906855266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,5120,0.04488889045185513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,1536,0.0542755557431115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,4096,0.037235554721620344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,4096,0.052758223480648465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,3584,0.03312355611059401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,4096,0.036551998721228704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,3584,0.04702755477693346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,768,0.03565511107444763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,3584,0.033075554503334895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,3072,0.04164177841610379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,3072,0.029659556017981634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,2560,0.02641600039270189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,2560,0.035679999325010515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,2560,0.026213333010673523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,2048,0.022297778063350256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,2048,0.02975822157329983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,65536,128,0.017314665847354464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,2048,0.022693332698610093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,1536,0.018744889232847426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,1536,0.02402488887310028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,1536,0.018530666828155518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,1024,0.016895999511082966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,768,0.013752000199423896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,1024,0.015511110424995422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,768,0.013176889055305056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,768,0.014074666632546319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,512,0.010111110905806223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,512,0.01310488912794325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,256,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,256,0.010689777632554373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,128,0.00906488878859414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,128,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,128,0.009574222067991892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,64,0.004800888813204235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,64,0.01015911085738076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,16384,32,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,16384,32,0.012219555675983429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,65536,0.3431546688079834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,10240,0.08732710944281684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,65536,0.5724444389343262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,8192,0.06887466377682157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,16384,0.14147022035386828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,16384,0.09659910864300197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,16384,0.11817333433363174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,12288,0.06959822442796496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,12288,0.10937866899702285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,12288,0.09021421935823228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,65536,0.5714249081081814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,10240,0.05792533026801216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,10240,0.09270222319497003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,5120,0.04607377780808342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,8192,0.07439021931754218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,10240,0.0794631110297309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,8192,0.05588089095221626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,7168,0.06584178076850043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,7168,0.049011554982927114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,3072,0.02940888868437873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,6144,0.0581244428952535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,6144,0.0446337792608473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,5120,0.03299644589424133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,5120,0.049646221929126315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,5120,0.036933332681655884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,4096,0.04002133342954848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,4096,0.031148443619410198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,1024,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,3584,0.03631466627120972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,512,0.01219911128282547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,3584,0.02816177739037408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,3072,0.0321688883834415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,3072,0.02309422194957733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,3072,0.025082666012975905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,16384,256,0.010341333018408881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,2560,0.02760355505678389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,2048,0.017864000466134813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,2560,0.02206666602028741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,2048,0.023039999935362074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,2048,0.019190222024917603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,1536,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,1536,0.018401778406567044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,1536,0.01590666671593984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,1024,0.012638222012254926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,1024,0.012863111164834766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,1024,0.013564444250530668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,768,0.010562666588359408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,768,0.012739555703269111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,512,0.010433778166770935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,512,0.008373333348168267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,512,0.011624000138706632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,256,0.00942666663063897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,256,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,256,0.009196444518036312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,128,0.007438222567240397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,128,0.0046568889584806235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,128,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,64,0.0041982221106688184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,64,0.008918222453859117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,12288,32,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,12288,32,0.010354667074150508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,65536,0.3121448887719048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,65536,0.47993511623806423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,65536,0.47578753365410703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,16384,0.08224266767501831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,16384,0.12626043955485025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,16384,0.09389866722954644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,12288,0.06289955642488268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,7168,0.042783111333847046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,8192,0.04725689027044508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,12288,0.0982053346104092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,6144,0.038182222180896334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,12288,0.07259288761350843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,10240,0.08160177866617839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,10240,0.06749422020382352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,4096,0.027449778384632532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,8192,0.07084000110626221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,8192,0.04548177785343594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,3584,0.025613332788149517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,7168,0.038359112209743924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,7168,0.05906666649712456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,7168,0.04073244333267212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,6144,0.0343795551194085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,6144,0.05173244410090976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,5120,0.04632177617814806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,6144,0.0373564428753323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,5120,0.02942844563060337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,4096,0.03528889020284017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,5120,0.031161778502994116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,4096,0.02615822189384037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,2560,0.02051111062367757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,3584,0.032249778509140015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,3584,0.02423733307255639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,3072,0.020298666424221463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,3072,0.0295724438296424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,3072,0.021242666575643752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,2560,0.02518133322397868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,2560,0.01928711103068458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,2048,0.016186666157510545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,2048,0.021148444877730474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,12288,768,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,2048,0.016773333152135212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,1536,0.014093332820468478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,1536,0.01740977830357022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,1536,0.014020444618331062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,1024,0.011671110987663269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,1024,0.012264000044928657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,1024,0.0114702218108707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,768,0.009770666559537252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,768,0.010444444086816577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,768,0.01144088887506061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,512,0.009308444129096137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,512,0.007753777835104201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,512,0.010201777848932479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,256,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,256,0.00852888905339771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,128,0.006664888726340399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,128,0.004568888909286923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,128,0.008055999875068665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,64,0.004071111066473855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,64,0.008420444197124904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,10240,32,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,10240,32,0.00962311112218433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,65536,0.4028462304009332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,65536,0.23731022410922578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,16384,0.06362311045328777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,16384,0.09876266453001235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,65536,0.35743199454413516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,16384,0.07078311178419325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,12288,0.07735911342832777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,12288,0.052397333913379245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,12288,0.05570489168167114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,10240,0.04411733481619093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,10240,0.0671608911620246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,10240,0.05148266752560934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,8192,0.035952889257007174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,8192,0.037232001622517906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,8192,0.05214222272237142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,7168,0.04707555638419258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,7168,0.03275555703375075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,7168,0.03308088911904229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,6144,0.028844445943832397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,6144,0.041702220837275185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,6144,0.029712888929578993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,5120,0.025033776958783466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,5120,0.040256887674331665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,5120,0.02606222199069129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,4096,0.028039998478359644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,4096,0.02124622298611535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,4096,0.02201244400607215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,3584,0.02527911133236355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,3584,0.020089776979552377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,3584,0.020232000284724765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,3072,0.01793866687350803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,3072,0.01797777745458815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,2560,0.016385777129067313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,3072,0.022879110442267522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,2560,0.01939733326435089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,2560,0.01627288924323188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,2048,0.014275555809338888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,2048,0.01592088904645708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,2048,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,1536,0.012416000167528788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,1536,0.012627555264366997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,1536,0.012303999728626676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,1024,0.009443555441167619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,1024,0.010696888797812991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,1024,0.011013333168294696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,768,0.009657777845859528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,768,0.008113777471913232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,768,0.010090666512648264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,512,0.008613333106040955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,512,0.006728888799746831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,512,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,256,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,256,0.004481777962711123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,256,0.0075733330514695905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,8192,128,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,128,0.003960000144110786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,128,0.007171555525726742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,64,0.0036071112586392295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,64,0.007426666716734569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,8192,32,0.003969777789380816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,8192,32,0.008343111309740279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,10240,0.05249955587916904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,65536,0.3486062155829535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,65536,0.3013422224256727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,8192,0.04283644424544441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,16384,0.07524000273810492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,16384,0.10488622056113349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,3584,0.022886221607526142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,16384,0.06573600239223905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,12288,0.06684444348017375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,10240,0.0469466679626041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,12288,0.05293155709902445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,10240,0.06962222523159452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,4096,0.024653333756658766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,8192,0.04551555713017782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,65536,0.31817600462171763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,10240,0.05189777745140923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,8192,0.03797422183884515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,8192,0.03663377629386054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,7168,0.03315555718210008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,7168,0.03402222196261088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,6144,0.02937600016593933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,7168,0.0406675570540958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,6144,0.03613600134849548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,5120,0.032412442896101214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,6144,0.029409776131312054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,5120,0.025794666674402025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,4096,0.025224000215530396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,4096,0.02105866703722212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,3584,0.02292622294690874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,4096,0.022230222821235657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,3584,0.019446222318543326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,3072,0.016926222377353244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,3072,0.020853334003024634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,3072,0.017433777451515198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,2560,0.01774666706720988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,2560,0.015111999379263984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,2560,0.018363555272420246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,2560,0.015799111790127225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,2048,0.013415111435784234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,2048,0.014474666780895658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,1536,0.011604444848166572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,2048,0.0140346669488483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,1024,0.01055111073785358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,1536,0.01218488895230823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,1024,0.008991111483838823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,1024,0.010944888823562197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,768,0.007819555699825287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,768,0.009344889058007134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,768,0.010098666780524785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,10240,256,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,512,0.008464888566070134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,512,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,512,0.008184889124499427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,256,0.0046880000995265115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,256,0.0075288886825243635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,128,0.0041831110914548235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,128,0.0058062221441004015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,64,0.00387999994887246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,128,0.007135111424658034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,7168,32,0.004122666600677702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,64,0.007491555478837755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,7168,32,0.008083555433485243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,65536,0.300111108356052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,65536,0.18457689550187853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,16384,0.05490666627883911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,65536,0.32009866502549916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,16384,0.08410933282640244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,16384,0.0665102203687032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,12288,0.05863200293646919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,12288,0.03920533259709676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,10240,0.06009778049257067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,10240,0.035713778601752386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,12288,0.051875554853015475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,8192,0.03937066594759623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,10240,0.04809866680039299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,8192,0.035071111387676664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,7168,0.025431111454963684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,7168,0.035516444179746844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,7168,0.03178400132391188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,6144,0.03325244453218248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,6144,0.028361777464548748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,5120,0.020046222541067336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,5120,0.030392888519499037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,5120,0.025126222107145522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,4096,0.01758399936887953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,4096,0.022522666388087805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,4096,0.020593777298927307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,3584,0.016427555018001132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,3584,0.02012711101108127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,3584,0.01872177753183577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,3072,0.014861333701345654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,3072,0.01865600049495697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,3072,0.01722666621208191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,2560,0.013637333280510373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,2560,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,2560,0.015568888849682279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,2048,0.01220355596807268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,2048,0.013736888766288757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,2048,0.01386844449573093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,1536,0.010975999964608086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,1536,0.010825778047243753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,1536,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,1024,0.010041777458455827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,1024,0.008103999826643202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,1024,0.010527110762066312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,768,0.008815999660227034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,768,0.00700800038046307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,768,0.008616888688670265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,12288,0.057075553470187716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,512,0.0057813334796163775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,512,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,256,0.005694222119119432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,256,0.0041519999504089355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,256,0.00738666703303655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,128,0.005371555685997009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,128,0.00369422220521503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,128,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,64,0.003399111123548614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,64,0.007186666958861881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,6144,32,0.003624888757864634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,6144,32,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,3584,0.01921066641807556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,5120,0.0251866661840015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,65536,0.2451039950052897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,1536,0.011873777541849347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,16384,0.0670880013042026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,16384,0.07369155353969999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,65536,0.23924265967475045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,16384,0.05096977618005541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,7168,256,0.0063573333124319715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,12288,0.05043022169007195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,12288,0.057024889522128634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,10240,0.04830577638414171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,12288,0.03914755582809448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,10240,0.043549332353803843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,10240,0.03524000114864773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,8192,0.033896889951494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,8192,0.03868800070550706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,8192,0.0269128895468182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,7168,0.03130133284462823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,7168,0.024492444263564214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,6144,0.030453334252039593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,6144,0.027310222387313843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,6144,0.022125333547592163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,5120,0.023196443915367126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,5120,0.025960889127519395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,4096,0.021503999829292297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,5120,0.01947377787695991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,4096,0.018773333893881906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,3584,0.019152889649073284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,4096,0.016594666573736403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,3584,0.01536799967288971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,3072,0.01496088835928175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,3072,0.01700177788734436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,3072,0.014008000493049622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,2560,0.014601776997248331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,2560,0.01293511109219657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,2048,0.011583999627166323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,2048,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,2048,0.011713777979214987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,8192,0.028239111105600994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,1536,0.010319111247857412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,1536,0.010401777923107147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,1024,0.008975110948085785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,1024,0.007337777978844113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,1024,0.008927110996511247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,6144,0.022634666826989915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,768,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,768,0.00647199981742435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,768,0.007684444387753804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,512,0.005697777701748743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,512,0.0064444442590077715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,512,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,256,0.005692444327804778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,256,0.004048000193304486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,256,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,128,0.005324444423119227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,128,0.0037519998020595978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,128,0.0063902222447925145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,64,0.0034497777620951333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,64,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,5120,32,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,5120,32,0.006950221955776215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,65536,0.19992532995012072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,6144,512,0.007257777783605788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,16384,0.04518399967087639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,16384,0.06531911426120333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,65536,0.2032515605290731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,16384,0.045773333973354764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,12288,0.03568177753024631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,12288,0.053247110711203686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,12288,0.03581155671013726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,10240,0.030429333448410034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,10240,0.051514665285746254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,10240,0.03138577938079834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,8192,0.024714666936132643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,8192,0.03156977891921997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,8192,0.026085333691702947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,7168,0.022516444325447083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,7168,0.032487998406092324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,7168,0.023892444041040208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,6144,0.02012977831893497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,6144,0.030089779032601252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,65536,0.2674497763315837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,6144,0.02095199955834283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,5120,0.017809778451919556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,5120,0.026036444637510512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,5120,0.01882133384545644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,4096,0.015340444114473132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,4096,0.016691555579503376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,4096,0.016072000066439312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,3584,0.015211555692884656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,3584,0.01463644372092353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,3072,0.013085333009560903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,3072,0.014112000664075216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,3072,0.013887999786270989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,2560,0.012079111403889127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,2560,0.011913778053389655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,2560,0.012694222231705984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,2048,0.010106667048401302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,2048,0.010524444282054901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,2048,0.011691555380821228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,1536,0.009462222456932068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,1536,0.008597333398130205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,1536,0.010141332944234213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,1024,0.007254222200976477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,7168,0.030639999442630347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,1024,0.008007999923494127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,1024,0.006721777634488211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,768,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,768,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,512,0.004488888714048597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,512,0.006834666762087081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,256,0.004879111217127906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,256,0.003801777958869934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,256,0.006471111128727595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,3584,0.017297777864668105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,128,0.003452444449067116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,128,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,64,0.0032586666444937387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,64,0.006509333435032103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,4096,32,0.0034373334298531213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,4096,32,0.006743999818960826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,2560,0.013439111411571503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,65536,0.1769031153784858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,65536,0.1553608841366238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,5120,1536,0.010118222071064843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,16384,0.043805334303114146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,16384,0.05824266539679634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,65536,0.19841244485643175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,16384,0.04597599969969856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,12288,0.035707556539111666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,12288,0.054720001088248364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,10240,0.03789777888192071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,10240,0.030972444348865088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,8192,0.02385866641998291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,8192,0.04061866799990336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,8192,0.025232889586024817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,65536,0.1602133380042182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,7168,0.027452444036801655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,7168,0.023528888821601868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,6144,0.01958488921324412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,6144,0.021244444780879553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,6144,0.020799110333124798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,3584,0.014319999350441827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,5120,0.018413333429230582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,5120,0.018922666708628338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,4096,0.015042665931913586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,4096,0.01497066683239407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,512,0.005419555637571547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,768,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,3584,0.013931555880440606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,4096,0.016024889217482675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,3584,0.01808177762561374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,3584,0.014961777461899651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,3072,0.012379555238617791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,3072,0.01275644451379776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,3072,0.013795555465751223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,2560,0.012475555141766867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,2560,0.012874667015340594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,2048,0.011128889189826118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,2048,0.011378666592968835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,1536,0.009144888983832465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,4096,128,0.0047146665553251905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,1536,0.008942222429646386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,1536,0.009538667069541084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,1024,0.006568888823191325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,1024,0.0069857777820693115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,1024,0.008089777496125963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,768,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,768,0.005453333258628845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,768,0.00738933351304796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,10240,0.029206222958034937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,12288,0.03428355521625943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,512,0.006864000111818314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,512,0.004423999951945411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,256,0.0037973332736227247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,256,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,256,0.006522666662931442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,128,0.004652444273233414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,128,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,128,0.0034728890491856467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,64,0.0032604444358083936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,64,0.006535111202134027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3584,32,0.00342399999499321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3584,32,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,7168,0.021906667285495337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,65536,0.1542631122801039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,16384,0.05509600043296814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,16384,0.0459173321723938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,65536,0.1714195542865329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,16384,0.04173688756095039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,12288,0.038556443320380315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,12288,0.0436462230152554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,12288,0.033041778537962176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,10240,0.037462221251593694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,10240,0.028804444604449805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,10240,0.029302222861184016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,8192,0.023392889234754775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,8192,0.02398577829202016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,7168,0.026386666629049513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,7168,0.02127555509408315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,7168,0.022055110997623865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,6144,0.023546665906906128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,6144,0.01884711119863722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,6144,0.01980977753798167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,5120,0.01980355547534095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,5120,0.021073778470357258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,5120,0.017882666654056974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,4096,0.016672000288963318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,4096,0.013456000222100152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,4096,0.015241778559154935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,5120,0.0174319992462794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,3584,0.01202222208182017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,3584,0.015133332875039844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,3584,0.01440088947614034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,3072,0.014355555176734924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,3072,0.01366222235891554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,3072,0.013330666555298699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,2560,0.012151111331250934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,2560,0.011679111255539788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,2048,0.009887999958462184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,2048,0.010406222608354356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,2048,0.010928000013033548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,1536,0.008195555872387355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,1536,0.009151111046473185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,2560,0.011694221860832639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,1024,0.006361777583758037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,1024,0.007619555625650618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,2048,0.010410666465759277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,768,0.005130666825506422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,768,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3584,512,0.0053226666318045715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,512,0.004294222427739037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,512,0.006802666518423293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,256,0.005160000175237656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,256,0.00378933341966735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,256,0.0064702220261096954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,65536,0.22983021206325957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,128,0.0034204444123639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,128,0.006206222292449739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,64,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,64,0.006433777925040986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,3072,32,0.0034008889148632684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,3072,32,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,8192,0.029551111989551123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,65536,0.09859822193781535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,65536,0.1350568930308024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,16384,0.02840177880393134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,65536,0.1149982213973999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,16384,0.04297066728274027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,12288,0.022407111194398668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,16384,0.029037333197063867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,12288,0.03267644511328803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,12288,0.02334755493534936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,10240,0.02529777752028571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,10240,0.019917334119478863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,10240,0.02067377832200792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,2560,0.01200799975130293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,8192,0.023203555080625746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,8192,0.01755288905567593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,1536,0.008848889006508721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,7168,0.0185208883550432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,7168,0.01646844380431705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,1024,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,6144,0.014007111390431723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,6144,0.018595554762416415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,6144,0.015077332655588785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,5120,0.012803555362754397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,5120,0.014340443743599786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,5120,0.013783111340469785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,4096,0.011680000358157687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,4096,0.01202400028705597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,3584,0.010791111323568555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,3584,0.01071733319097095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,3584,0.011438222395049201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,3072,0.011046222514576383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,3072,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,2560,0.009516444471147325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,2560,0.009769777456919352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,2560,0.010090666512648264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,768,0.006184000107977126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,2048,0.007537778053018782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,2048,0.009001777403884465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,512,0.005665777872006099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,1536,0.007139555282062954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,1536,0.007365333537260692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,3072,128,0.004839111119508743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,1024,0.005842666659090254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,1024,0.006597333484225803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,768,0.005409777578380372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,768,0.00435644429590967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,768,0.00637511122557852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,512,0.00507377791735861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,512,0.003818666769398583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,512,0.006137777947717243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,256,0.004757333132955763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,256,0.0035706667436493766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,256,0.0058257778485616045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,128,0.004506666627195147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,128,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,128,0.005653333332803514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,64,0.003006222140457895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,64,0.0058328890138202244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2560,32,0.0030906666070222855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2560,32,0.00601333338353369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,65536,0.09204266468683879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,65536,0.11014399925867717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,65536,0.10082933637830947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,16384,0.026851556367344324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,16384,0.03992977738380432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,16384,0.027649778458807204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,12288,0.021386666430367365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,12288,0.032968888680140175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,12288,0.02254666719171736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,10240,0.018958222534921434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,10240,0.028535998529858057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,10240,0.01975733372900221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,8192,0.015812445018026564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,8192,0.01904177831278907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,8192,0.017098666893111337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,8192,0.016873778568373788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,7168,0.017161778277821012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,7168,0.015541333291265698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,7168,0.015863110621770222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,6144,0.0151555554734336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,6144,0.01442577772670322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,4096,0.01129066695769628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,5120,0.017688888642523024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,5120,0.013334222137928009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,4096,0.010854221880435944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,3072,0.010089777410030365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,4096,0.01460800071557363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,4096,0.011662222445011139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,3584,0.010318222145239512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,3584,0.01109333336353302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,3584,0.01295111080010732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,3072,0.011597333682907952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,3072,0.009655999640623728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,3072,0.010424888796276517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,2560,0.008919999831252629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,2560,0.009887999958462184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,2560,0.009484444227483537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,2048,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,2048,0.008754666480753157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,2048,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,1536,0.006455111006895701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,1536,0.00721688899728987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,1536,0.00701155596309238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,1024,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,1024,0.006521777974234686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,768,0.005383111122581694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,768,0.004861333303981357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,768,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,512,0.005105777747101254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,512,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,512,0.006091555373536215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,256,0.0047280001971456735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,256,0.0033884443756606844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,256,0.0057760001056724125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,2048,0.00868800034125646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,128,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,128,0.005667555663320754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,64,0.003006222140457895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,1536,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,64,0.005731555736727185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,2048,32,0.0031724443866146933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,2048,32,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2560,1024,0.005807999935415056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,65536,0.08632977803548177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,65536,0.08688444561428493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,16384,0.030107554462220933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,16384,0.025813332862324182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,7168,0.014768888552983602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,12288,0.024527110987239416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,12288,0.02085866696304745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,10240,0.01880088945229848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,10240,0.022072000635994807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,10240,0.01872711049185859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,6144,0.013408000270525614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,8192,0.018339556124475267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,8192,0.01623911162217458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,5120,0.0120666664507654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,7168,0.01642311116059621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,7168,0.015207999282413058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,6144,0.013170666992664337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,6144,0.01331822243001726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,6144,0.01391999920209249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,5120,0.011908444265524546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,5120,0.012915555801656512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,5120,0.012833777401182385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,1024,0.0057591112951437635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,4096,0.010096888575288985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,4096,0.011179555621412067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,3584,0.010079111489984725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,3584,0.010072888599501716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,3584,0.010752889017264048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,3072,0.009112888740168678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,3072,0.009172444542249044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,2048,128,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,3072,0.009770666559537252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,2560,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,2560,0.008062221937709385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,2560,0.008546666966544257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,2048,0.007152000235186682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,2048,0.007744889292452071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,1536,0.00637511122557852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,1536,0.00608800020482805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,1536,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,65536,0.08948266506195068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,1024,0.004781333522664176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,1024,0.0064737776087390045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,16384,0.026575111680560645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,768,0.0041706665522522396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,768,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,512,0.004953777624501122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,512,0.0036888888312710654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,512,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,12288,0.02109066645304362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,256,0.003354666754603386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,256,0.0057839999596277876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,7168,0.014549333188268872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,128,0.0030906666070222855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,128,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,8192,0.015582222077581616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,64,0.0028986665937635633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1536,32,0.002983999955985281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,64,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1536,32,0.0058062221441004015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,65536,0.06267199913660686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,65536,0.16926133632659912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,65536,0.08466133144166733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,16384,0.043712000052134194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,4096,0.010615111225181155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,16384,0.024188444018363953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,16384,0.024918221765094336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,12288,0.02018311123053233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,12288,0.03436177637841966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,12288,0.02018044392267863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,10240,0.0181048893266254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,10240,0.016227554943826463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,8192,0.013897778259383308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,8192,0.023601777023739282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,8192,0.015612444943851896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,7168,0.02102488941616482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,7168,0.01311377767059538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,7168,0.014621333943472968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,6144,0.012279999752839407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,6144,0.013563555147912769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,5120,0.016005333926942613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,5120,0.010752889017264048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,5120,0.012601777911186218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,4096,0.012917333179050021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,4096,0.009365333451165093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,4096,0.010808000134097205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,2048,0.007131555842028723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,3584,0.008647999829716152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,3584,0.009635555247465769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,3072,0.01032977799574534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,3072,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,1024,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,3072,0.008866666919655269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,2560,0.009327111144860586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,2560,0.007888000044557784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,2560,0.007461333440409766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,2048,0.008192000289758047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,2048,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,2048,0.007329777710967594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,768,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,1536,0.005161777966552311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,1536,0.006830222076839871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,256,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,1024,0.004297777596447202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,1024,0.006400888992680444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,768,0.003995555556482739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,768,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,512,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,512,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,512,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,256,0.0047306666771570844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,256,0.0031911111954185697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,256,0.005680000202523337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1536,128,0.004503111044565837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,128,0.002975111206372579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,128,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,64,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,64,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,1024,32,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,1024,32,0.005668444352017508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,6144,0.018539556198649935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,65536,0.04991199904017978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,10240,0.02959022257063124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,65536,0.08203466733296712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,16384,0.017458667357762653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,16384,0.03969688879119025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,16384,0.02472355630662706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,12288,0.03195733494228787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,12288,0.014538667268223233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,12288,0.01994400057527754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,10240,0.015495111544926962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,10240,0.02747111188040839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,10240,0.01758044461409251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,8192,0.022738666998015508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,3584,0.011576000187132092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,8192,0.012864000267452665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,8192,0.015447111593352424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,7168,0.013623110949993134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,7168,0.014463999205165438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,6144,0.012011555333932241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,6144,0.013376889129479727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,5120,0.015486222174432544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,5120,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,5120,0.012313777373896705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,4096,0.012664889295895895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,4096,0.00926577796538671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,4096,0.010311110979980892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,3584,0.011329777538776398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,3584,0.008608889248636033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,3584,0.009363555245929295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,1536,0.007287999822033777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,3072,0.007869333028793335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,3072,0.008387555678685507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,2560,0.00921688891119427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,2560,0.007112888826264276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,2560,0.007807110746701558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,2048,0.007887110941939885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,2048,0.005809777726729711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,2048,0.007294221884674496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,1536,0.00685511115524504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,1536,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,1536,0.0068195557428730865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,1024,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,1024,0.004109333372778363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,1024,0.006437333507670297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,768,0.005436444448100195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,768,0.00378577783703804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,768,0.006210666563775804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,512,0.005116444494989184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,512,0.0034382223255104492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,512,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,1024,0.006148444281684027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,256,0.003093333293994268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,256,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,768,0.005647999958859549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,128,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,128,0.005468444277842839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,64,0.0026853332916895547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,64,0.005533333453867171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,768,32,0.002708444371819496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,768,32,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,65536,0.07701777749591403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,65536,0.0387644436624315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,1024,128,0.004468444320890638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,65536,0.07801955276065402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,16384,0.0140364451540841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,16384,0.02349600030316247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,12288,0.018294221825069852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,12288,0.013592888911565145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,12288,0.01930488811598884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,10240,0.016169778174824186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,10240,0.01257155587275823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,10240,0.017496888836224873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,8192,0.013582222163677216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,8192,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,65536,0.14029688305324978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,8192,0.0151591118839052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,7168,0.010576888918876648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,7168,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,6144,0.01091466678513421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,6144,0.009999111294746399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,6144,0.012974222501118978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,6144,0.018384888768196106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,7168,0.020706666840447318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,5120,0.011551111108726926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,5120,0.009511999785900116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,4096,0.008692444198661381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,4096,0.008641777767075432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,4096,0.009463110731707679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,3584,0.007887110941939885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,3584,0.008829333715968663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,3584,0.008279110822412703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,3072,0.007374222079912822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,3072,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,3072,0.008368000388145447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,2560,0.006849777781301075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,2560,0.006252444452709622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,2560,0.007778666913509369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,3072,0.01012711144155926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,2048,0.005515555540720622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,2048,0.006319111006127463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,2048,0.00721244431204266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,1536,0.004834666848182678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,1536,0.005789333333571752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,1536,0.006780444333950679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,1024,0.004108444270160464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,1024,0.0053013331360287136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,1024,0.0063457778758472866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,768,0.0037786666717794207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,768,0.006176888942718506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,512,0.0034657776769664553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,512,0.0058471109304163195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,256,0.00443111111720403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,256,0.003097777772280905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,256,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,128,0.004271111140648524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,128,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,128,0.00546577779783143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,64,0.002696000039577484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,64,0.005495999836259418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,512,32,0.0027315556589100095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,512,32,0.005519111123349931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,65536,0.07768088579177856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,65536,0.02968533171547784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,65536,0.07779644595252143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,16384,0.022671999202834234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,16384,0.013545778062608508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,16384,0.023511999183230933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,12288,0.01827911039193471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,12288,0.011588444312413534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,12288,0.019392000304328073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,10240,0.015944000747468736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,10240,0.010656889114114972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,10240,0.01738577749994066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,8192,0.013025778035322825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,8192,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,8192,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,7168,0.011830222275522021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,7168,0.00924177798959944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,7168,0.013070222404268054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,6144,0.010335110955768162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,6144,0.008607111043400234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,6144,0.011932444241311816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,256,0.004654222064548069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,5120,0.00926399976015091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,5120,0.0075457774930530125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,5120,0.010337777435779572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,4096,0.006791111081838608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,4096,0.009165333376990424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,3584,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,3584,0.0070497774415545994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,3584,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,3072,0.007191999918884701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,3072,0.006700444552633498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,3072,0.008231110870838165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,2560,0.006666666517655055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,2560,0.006141333530346553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,2560,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,2048,0.006170666466156642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,2048,0.005420444326268301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,2048,0.007189333438873291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,1536,0.005685333162546158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,1536,0.0047884442740016514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,1536,0.006709333509206772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,1024,0.0051893335249688895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,1024,0.004078222231732475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,1024,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,768,0.004941333499219683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,768,0.0037742224004533556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,768,128,0.004416000097990036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,768,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,512,0.003452444449067116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,512,0.004613333278232151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,512,0.0057946667075157166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,256,0.003060444361633725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,16384,0.022853333089086745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,128,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,128,0.005475555443101459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,64,0.0027377777215507296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,64,0.005476444545719359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,256,32,0.002712888850106133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,256,32,0.005487111293607288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,65536,0.0773271123568217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,65536,0.023504000571038988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,65536,0.07715466949674818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,16384,0.022488888767030504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,16384,0.01018044435315662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,16384,0.023451555106374953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,12288,0.017924444542990792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,12288,0.009104889300134447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,12288,0.0192248887485928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,10240,0.015282667345470853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,10240,0.008607111043400234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,10240,0.017065778374671936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,8192,0.01276088919904497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,8192,0.007730666961934831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,8192,0.013989332649442883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,7168,0.011586666935020022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,7168,0.006719111154476802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,7168,0.012331555287043253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,6144,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,6144,0.006544000158707301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,7168,0.01218577805492613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,6144,0.011224889092975192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,5120,0.00924711094962226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,5120,0.006832888970772426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,5120,0.010303110712104374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,4096,0.008202667037645975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,4096,0.009129777550697327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,4096,0.006412444429265128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,3584,0.006812444577614467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,3584,0.007717333734035492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,3584,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,3072,0.007190222541491191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,3072,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,3072,0.008207110895050896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,2560,0.006024000131421619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,2560,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,2048,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,2048,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,2048,0.007114667031500075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,1536,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,1536,0.004616888860861461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,1536,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,1024,0.005076444397370021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,5120,0.009647111097971598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,1024,0.003959111041492886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,1024,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,768,0.004837333328194088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,768,0.00361333332127995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,768,0.006040888941950268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,512,0.0033200000309281876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,512,0.005746666755941179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,256,0.004287999951177173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,256,0.003000000077817175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,256,0.005518222020732031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,128,0.004150222159094281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,128,0.0027928888383838865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,128,0.005427555491526921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,64,0.002620444529586368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,64,0.005437333136796951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,128,32,0.002595555658141772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,128,32,0.005421333428886201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,65536,0.01902044481701321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,65536,0.07645955350663927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,16384,0.008756444685988957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,16384,0.02329155637158288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,12288,0.007552000383536021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,12288,0.01891911029815674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,10240,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,10240,0.016514667206340365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,8192,0.00674311113026407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,8192,0.01293333371480306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,7168,0.006542222367392645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,7168,0.011945777469211154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,6144,0.0063902222447925145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,6144,0.010996444357766045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,5120,0.006735111276308696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,5120,0.01015644437736935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,4096,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,4096,0.009106666677527958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,3584,0.006792888873153263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,3584,0.008593777815500895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,3072,0.006463999963468975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,3072,0.008176888856622908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,2560,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,2560,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,2048,0.005281777845488654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,2048,0.007082666787836287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,1536,0.004629333400064045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,1536,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,1024,0.003931555483076307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,1024,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,768,0.003579555700222651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,768,0.005035555611054103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,768,0.006005333529578314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,512,512,0.0047111109726958806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,512,0.0058000000814596815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,512,0.0032799999333090256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,256,0.0029475556479560006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,128,0.0027662221756246355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,128,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,64,0.002567111204067866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,64,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,64,32,0.00258933338854048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,64,32,0.005482666608360078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,65536,0.07674488756391737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,16384,0.008235555556085374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,16384,0.023222222924232483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,65536,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,12288,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,12288,0.018783999813927543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,10240,0.0070168889231152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,10240,0.01644977761639489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,8192,0.006681777950790193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,8192,0.013027555412716336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,7168,0.006508444332414203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,7168,0.011940444509188334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,6144,0.0063493334584765965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,6144,0.010878221856223213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,5120,0.006729777902364731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,4096,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,5120,0.010082667072614035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,4096,0.008956444760163626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,3584,0.006759111252095964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,3584,0.0085235552655326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,3072,0.006395555618736479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,3072,0.00813155538505978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,2560,0.0060159998635451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,2560,0.007628444168302748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,2048,0.005292444593376583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,2048,0.007075555622577667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,1536,0.004641777939266629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,1536,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,1024,0.003961777935425441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,1024,0.006183111005359226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,768,0.003593777616818746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,768,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,512,0.0032808888289663526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,512,0.005704000178310607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,256,0.0029360000044107437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,256,0.005479111025730769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,128,0.002732444347606765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,128,0.005336888962321811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,64,0.0025840000145965153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,64,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,192,32,32,0.002598222138153182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,192,32,32,0.0053546664615472155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,16384,0.7419030931260852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,4096,0.008259555531872643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,16384,0.8089111116197375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,256,0.004335111214054955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,12288,0.43640976481967503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,12288,0.567380428314209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,10240,0.4721653196546767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,12288,0.6207928657531738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,10240,0.5175075531005859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,8192,0.30904799037509495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,256,128,0.0041519999504089355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,8192,0.37988000445895725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,7168,0.33188266224331325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,7168,0.27201599544949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,8192,0.44028356340196395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,6144,0.239990234375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,6144,0.28844356536865234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,7168,0.390155553817749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,5120,0.24162311024136016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,6144,0.3222213321261936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,5120,0.272436433368259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,4096,0.19773422347174752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,4096,0.15375378396775988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,3584,0.17302044232686362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,3584,0.13832888338300917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,4096,0.21326043870713976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,3584,0.18248355388641357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,3072,0.15080266528659397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,3072,0.11539467175801595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,2560,0.127454227871365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,2560,0.09783822298049927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,3072,0.16377422544691297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,2048,0.10483821895387437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,2048,0.08041866620381673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,2560,0.006712889091836081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,2560,0.13807822598351374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,2048,0.10847555266486274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,192,128,512,0.004512000001139111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,1536,0.05935111310746935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,1536,0.08209511306550768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,1024,0.05831466780768501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,1536,0.08297599686516656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,1024,0.04075555668936835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,1024,0.05641511413786146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,768,0.046613332298066884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,512,0.03487199876043532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,768,0.0413315561082628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,256,0.020230222079488967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,512,0.03527733352449205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,256,0.021703110800849065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,256,0.03127911024623447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,128,0.015698666373888653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,128,0.013271999855836233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,16384,0.5551502439710829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,128,0.02834222382969326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,64,0.010630221830474006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,65536,32,0.01015822258260515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,32,0.03829688827196757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,65536,64,0.03043022089534336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,65536,0.7618213229709201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,10240,0.37703198856777614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,65536,0.5471893416510688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,16384,0.19133333365122476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,65536,0.9123146269056531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,12288,0.12005778153737386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,16384,0.21426755852169463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,12288,0.13999556170569524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,10240,0.11826044983334011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,12288,0.1560008923212687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,10240,0.12876799371507433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,8192,0.09770399994320339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,8192,0.07716977596282959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,7168,0.08359822299745347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,7168,0.06937066713968913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,8192,0.11015733083089192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,6144,0.07360711362626818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,7168,0.09555644459194607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,6144,0.08149422539605035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,5120,0.05134310987260607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,5120,0.0621617767545912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,5120,0.07080800003475614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,4096,0.041604446040259466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,4096,0.051163554191589355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,3584,0.03757155603832669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,4096,0.0539991127120124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,3584,0.04565511147181193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,3584,0.04642400145530701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,3072,0.03218488891919454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,3072,0.04043111205101013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,5120,0.19569777117835152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,3072,0.04222044348716736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,2560,0.027650665905740526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,2560,0.03566044569015503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,2560,0.03910933269394769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,2048,0.023678221636348303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,2048,0.029071112473805744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,1536,0.02334044377009074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,2048,0.029410666889614526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,1536,0.019824888971116807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,1024,0.016399999459584553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,1536,0.023553777072164748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,1024,0.01794488893614875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,768,0.01275022245115704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,768,0.01517955627706316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,512,0.01185955521133211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,512,0.009697777529557547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,512,0.013177778157922955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,256,0.010142222046852112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,256,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,256,0.010233778092596266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,128,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,128,0.00517155561182234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,128,0.009512888888518015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,64,0.004596444467703502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,64,0.010185778141021729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,16384,32,0.00500266667869356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,16384,32,0.012784889174832238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,512,0.026327111654811438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,65536,768,0.0344524449772305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,65536,0.39661243226793075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,65536,0.5646870930989584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,16384,0.10261333651012844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,16384,0.16183377636803523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,16384,0.16021422545115152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,12288,0.12386222680409749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,65536,0.6668142212761773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,12288,0.12012444602118598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,16384,0.1491911146375868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,10240,0.09084800216886733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,10240,0.10313600301742554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,8192,0.05117244521776835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,8192,0.07329777876536052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,8192,0.08509688907199436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,7168,0.04545333319240146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,7168,0.0657822224828932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,7168,0.07552355527877808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,6144,0.04020977682537503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,6144,0.05762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,10240,0.09977955288357204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,5120,0.03370399938689338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,6144,0.06521688567267524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,5120,0.04897866646448771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,5120,0.05630577935112847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,4096,0.028391109572516546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,4096,0.03940177626079983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,3584,0.025943999489148457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,4096,0.042661332421832614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,3584,0.035190221336152815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,3584,0.03800977932082282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,3072,0.023355555203225877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,3072,0.0311155551009708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,6144,0.061794665124681264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,2560,0.020759999752044678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,3072,0.03440355592303806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,2560,0.026877333720525105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,2560,0.02881244487232632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,2048,0.022484444909625586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,2048,0.023847111397319373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,1536,0.015672889020707872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,1536,0.017811555001470778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,1536,0.01921333372592926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,1024,0.012881778180599213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,1024,0.01238044434123569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,1024,0.015321777926550971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,768,0.010297777752081553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,768,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,512,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,512,0.008214222060309516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,512,0.01187111106183794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,256,0.008775999976529015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,256,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,256,0.00925422211488088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,128,0.006873777757088344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,128,0.004277333203289243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,128,0.008736888567606607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,64,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,64,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,12288,32,0.004229333251714706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,12288,32,0.010876444478829702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,768,0.013448889056841532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,16384,1024,0.01520088811715444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,65536,0.49306668175591367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,65536,0.3663457764519586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,16384,0.14275644885169134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,16384,0.09420000182257758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,12288,0.07190489106708102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,16384,0.1360444492763943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,65536,0.5649573538038466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,12288,0.10103288624021743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,10240,0.060080885887145996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,12288,0.10344355636172825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,10240,0.08637599812613593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,8192,0.04723822077115377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,10240,0.08649244573381211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,8192,0.06419733497831556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,7168,0.057188444667392306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,7168,0.04245866669548882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,8192,0.07288710938559638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,6144,0.050557331906424634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,7168,0.060942219363318555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,6144,0.03786399960517883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,6144,0.05239911211861504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,5120,0.04275288846757677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,5120,0.031686223215527005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,5120,0.046592000457975596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,4096,0.03365777929623922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,4096,0.026120889517996047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,3584,0.03706311186154684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,4096,0.03534222311443753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,3584,0.023576888773176406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,3584,0.031383110417260066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,3072,0.029014223151736792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,3072,0.027452444036801655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,2560,0.018963555494944256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,2560,0.023504889673656885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,2560,0.02426311042573717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,2048,0.01977066695690155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,2048,0.020749333831999037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,1536,0.014031999640994601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,1536,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,1536,0.01702044407526652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,1024,0.011218667030334473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,1024,0.011544889046086205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,1024,0.013449778159459433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,768,0.010279110736317104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,768,0.009477333062224919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,768,0.012332444389661154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,512,0.009413333402739631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,512,0.007737778127193451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,512,0.010536889235178629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,256,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,256,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,256,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,128,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,128,0.004383111165629493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,128,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,64,0.0038888889054457345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,12288,0.07574399974611071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,10240,32,0.004071111066473855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,64,0.008290666672918532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,10240,32,0.010275555153687796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,65536,0.3937324417961969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,16384,0.07149155272377862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,65536,0.23343377643161348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,10240,0.06368622514936659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,16384,0.09348355399237739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,65536,0.43585067325168186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,16384,0.10460888677173191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,12288,0.04974755644798279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,12288,0.07247289021809895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,10240,0.06120622158050537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,10240,0.04271022147602505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,12288,0.0778577791319953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,8192,0.04939111073811849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,8192,0.035419556829664446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,10240,0.06668355729844835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,7168,0.04386666748258802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,8192,0.05481422278616163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,7168,0.032032890452278986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,6144,0.039260443713929914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,7168,0.04847288793987698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,6144,0.02821777926550971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,5120,0.037880887587865196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,6144,0.04208533300293816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,5120,0.03576177689764235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,4096,0.020786666207843356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,4096,0.02749600013097127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,4096,0.028200000524520874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,3584,0.01921688848071628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,3584,0.02460711035463545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,3584,0.025337777203983728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,3072,0.01753422286775377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,3072,0.02180800007449256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,2560,0.01588088936275906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,3072,0.02273955610063341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,2560,0.01886133352915446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,2560,0.020185778538386028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,2048,0.013800000150998434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,2048,0.015636444091796875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,2048,0.01794488893614875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,2048,0.016551110479566786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,1536,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,1536,0.014057776994175382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,1536,0.012285333540704517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,1024,0.01051644484202067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,1024,0.00924711094962226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,1024,0.011701333026091257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,768,0.009515555368529426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,768,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,768,0.010579555398888059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,512,0.006502222269773483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,512,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,512,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,12288,768,0.011698666546079846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,256,0.004297777596447202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,256,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,256,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,128,0.0038008888562520346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,64,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,128,0.007055111229419708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,64,0.00755555596616533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,8192,32,0.003867555409669876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,8192,32,0.008451555338170793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,65536,0.3406426641676161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,16384,0.08257777823342217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,16384,0.08374755250083075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,65536,0.3919066588083903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,16384,0.09945866796705459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,12288,0.06781333022647433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,12288,0.07373155487908258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,10240,0.05511200096872118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,10240,0.05448355608516269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,10240,0.06463199853897095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,8192,0.04390133420626322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,8192,0.05265955461396111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,7168,0.038620445463392473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,7168,0.03943199912707011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,7168,0.0464826656712426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,6144,0.03423466616206699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,6144,0.03525955478350321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,6144,0.03961155480808682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,5120,0.02843733297453986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,5120,0.029863999949561224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,5120,0.03388266762097677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,3072,0.02150933278931512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,4096,0.024637333220905725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,4096,0.026752889156341553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,10240,2048,0.016407999727461074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,3584,0.022310222188631695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,3584,0.025088888075616624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,3072,0.018681777848137748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,3072,0.01979733341270023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,3072,0.02197066611713833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,2560,0.016231111354298063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,2560,0.01705777810679542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,2560,0.019000000423855253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,2048,0.01367555558681488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,2048,0.01387555566098955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,2048,0.016128889388508268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,5120,0.024497777223587036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,1536,0.011201777391963534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,1536,0.01367111090156767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,1024,0.010372444159454769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,1024,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,1024,0.011523555550310345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,8192,128,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,768,0.009147555463843876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,768,0.007483555210961237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,768,0.01015822258260515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,512,0.007824000385072496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,512,0.006227555374304454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,512,0.008652444514963362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,12288,0.06449600060780843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,256,0.004291555533806483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,256,0.007520889242490132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,65536,0.32625865936279297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,128,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,128,0.007122666471534305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,128,0.0037902221083641052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,64,0.0036435553597079385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,64,0.007338666253619724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,7168,32,0.0037324445115195382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,7168,32,0.007889777421951294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,65536,0.2928684552510579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,8192,0.04298933347066244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,16384,0.07348889112472534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,16384,0.04913333389494154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,16384,0.09002577596240574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,12288,0.0580622222688463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,12288,0.06975199778874715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,65536,0.3884853257073297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,10240,0.05572444200515747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,10240,0.0602746672100491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,4096,0.023253333237436082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,8192,0.03935733437538147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,8192,0.028062221076753404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,7168,0.02565333412753211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,8192,0.04864355590608385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,7168,0.03568711214595371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,6144,0.022655111220147874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,7168,0.042912890513738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,6144,0.031726221243540444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,6144,0.038697779178619385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,5120,0.019902222686343722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,5120,0.02703999976317088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,5120,0.03201599915822347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,3584,0.021454221672481958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,4096,0.017343110508388944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,4096,0.02197511163022783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,4096,0.025867554876539443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,3584,0.020082667469978333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,3584,0.016202666693263583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,3584,0.02257599929968516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,3072,0.014729777971903482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,3072,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,2560,0.015188443991872998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,3072,0.02055911057525211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,2560,0.01354311075475481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,2048,0.012658666405412885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,2560,0.018784888916545443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,2048,0.01531644496652815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,1536,0.010885333021481832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,1536,0.01052177780204349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,1536,0.013140444126394061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,1024,0.00998577806684706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,1024,0.008266666697131263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,1024,0.010966222319338055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,768,0.007182222273614671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,768,0.008446222378147973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,768,0.009715555442704095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,512,0.006892444358931647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,512,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,512,0.008255999949243333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,256,0.005646222167544895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,256,0.004171555654870139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,256,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,128,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,128,0.003738666574160258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,128,0.007090667055712805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,64,0.007248000138335758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,64,0.0034719999465677473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,6144,32,0.008394666843944127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,6144,32,0.003580444388919406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,1536,0.011888888974984487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,65536,0.23912978172302246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,16384,0.07393955522113375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,65536,0.2896764543321398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,16384,0.07187377744250827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,12288,0.05719111363093058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,16384,0.07072177860471937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,65536,0.2916666666666667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,12288,0.057766218980153404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,12288,0.05407377746370104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,10240,0.04315200116899279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,10240,0.04883289006021288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,10240,0.04632711079385546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,8192,0.03501777847607931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,8192,0.03792622354295518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,7168,0.03165955675972833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,8192,0.03866488734881083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,7168,0.03455288873778449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,6144,0.02808977829085456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,7168,0.03296355406443278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,6144,0.03013333347108629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,6144,0.028397334946526423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,5120,0.025402666793929204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,5120,0.024762666887707178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,7168,256,0.006129777679840724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,4096,0.019881778293185763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,4096,0.021115556359291077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,3584,0.01864177816443973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,3584,0.018046221799320646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,3584,0.018619555566045973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,3072,0.0159244438012441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,3072,0.016569778323173523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,2560,0.014782221780882942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,2560,0.013726222018400827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,2560,0.01499288943078783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,65536,0.18330932988060844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,2048,0.012320000264379712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,2048,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,2048,0.013154666456911298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,1536,0.009853333234786987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,1536,0.01055111073785358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,1536,0.011307555768224927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,12288,0.03867022196451823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,1024,0.007889777421951294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,1024,0.009273777405420939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,768,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,768,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,768,0.008267555799749162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,512,0.0052968888647026485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,512,0.007148444652557373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,256,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,256,0.0041333333485656315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,256,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,128,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,128,0.0036728887094391715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,128,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,64,0.0033155555526415506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,64,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,5120,32,0.0033760000434186724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,5120,32,0.006844444407357111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,10240,0.03341689043574863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,65536,0.19375110997094047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,65536,0.16140711307525635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,16384,0.04494844542609321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,16384,0.056652446587880455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,16384,0.06890666484832764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,65536,0.2786053286658393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,6144,2048,0.012137778103351593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,12288,0.035290665096706815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,12288,0.04646044307284885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,10240,0.040192888842688665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,12288,0.04948711064126757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,10240,0.04219377703136868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,8192,0.024799111816618178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,8192,0.034006221426857844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,8192,0.03423377871513367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,7168,0.03132088979085287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,7168,0.03158311049143473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,6144,0.0200284438000785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,6144,0.02933244572745429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,6144,0.027504000398847792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,5120,0.025516443782382544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,4096,0.020672889219390023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,5120,0.025046222739749487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,5120,0.02381244467364417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,4096,0.01538044379817115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,4096,0.019186667270130582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,4096,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,3584,0.014316444595654806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,3584,0.014195554786258273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,3584,0.017518222332000732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,3072,0.012992888689041138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,3072,0.01575555569595761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,2560,0.011994666523403592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,2560,0.011199110911952125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,2560,0.0140337778462304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,3072,0.016549333930015564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,2048,0.00960000024901496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,2048,0.011872889267073737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,1536,0.009400000174840292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,1536,0.008170666793982187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,1536,0.01054222219520145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,1024,0.00887111077706019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,1024,0.006450666735569636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,1024,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,5120,512,0.0063768890168931745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,768,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,768,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,768,0.007793777518802219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,512,0.004373333520359463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,512,0.007092444433106317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,10240,0.029998222986857098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,256,0.0037048889531029593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,256,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,128,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,128,0.003350222276316749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,128,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,7168,0.022523555490705702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,64,0.003118222165438864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,64,0.006559111177921295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,4096,32,0.0034391110142072043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,4096,32,0.006723555425802867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,5120,0.01777777738041348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,65536,0.17327111297183564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,65536,0.2532426781124539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,3072,0.01297155519326528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,16384,0.05083555645412869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,2048,0.010604444477293225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,16384,0.06414755847718981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,12288,0.041667554113599986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,12288,0.033999999364217125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,10240,0.028745777077145044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,10240,0.036186665296554565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,12288,0.04936088787184822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,8192,0.024874667326609295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,10240,0.04085510969161987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,8192,0.03401155604256524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,7168,0.021627555290857952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,7168,0.022629333866967097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,7168,0.03056266572740343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,6144,0.01940355532699161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,6144,0.0203448881705602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,6144,0.026425777210129633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,5120,0.01722133325205909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,5120,0.017338666650984023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,5120,0.022631110416518316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,4096,0.014901333385043673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,4096,0.01391999920209249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,4096,0.01899111105336083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,1024,0.007335111498832703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,3584,0.013788444300492605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,3584,0.012721777790122561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,3072,0.011384888655609555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,3584,0.016728000508414376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,3072,0.015164444843928019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,2560,0.011501332951916588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,2560,0.010075555907355415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,2560,0.013535111314720578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,2048,0.008624888956546783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,2048,0.010234666367371878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,2048,0.0118568887313207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,1536,0.0073928890956772705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,1536,0.010303110712104374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,1024,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,1024,0.0058417779703934984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,1024,0.008223111430803934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,768,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,768,0.004671111288997862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,768,0.00776533368561003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,512,0.005335111171007156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,512,0.003984000119898054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,512,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,256,0.004906666775544484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,256,0.0035111109415690103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,256,0.00655644428398874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,512,0.005392888767851724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,128,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,128,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,4096,256,0.004865777575307422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,64,0.0032675556010670135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,64,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3584,32,0.0033413333197434745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3584,32,0.006714666469229593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,65536,0.15213333235846624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,65536,0.2503146595425076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,65536,0.20655200216505262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,16384,0.04515733321507772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,16384,0.050833778248892895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,65536,0.15565244356791177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,12288,0.03736266824934218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,12288,0.04019644525316026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,10240,0.04166222280926175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,10240,0.03255555695957608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,16384,0.043134222428003945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,10240,0.03455111053254869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,8192,0.03209599852561951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,8192,0.03231199913554721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,8192,0.029304001066419814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,7168,0.025791111919615004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,7168,0.025634666283925373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,8192,0.023839111129442852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,6144,0.02128177715672387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,6144,0.022759111391173467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,5120,0.0214675549003813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,5120,0.018053332964579265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,5120,0.019732443822754752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,3072,0.012542222109105853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,4096,0.01773244473669264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,4096,0.01576622161600325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,4096,0.01686844395266639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,3584,0.016012445092201233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,3584,0.014703111516104804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,3584,0.015308444698651632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,3072,0.014379555980364481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,3072,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,3072,0.014081777797804939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,2560,0.009544000029563904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,2560,0.012759110993809171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,2048,0.011207111179828644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,1536,0.008974221845467886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,2048,0.008258666429254744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,2048,0.011092444260915121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3584,128,0.0046453335218959385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,1536,0.007262222468852997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,1536,0.00960799968904919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,1024,0.00701333334048589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,1024,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,1024,0.008031110796663497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,768,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,768,0.004592000196377437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,768,0.00755466686354743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,512,0.005769777629110549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,512,0.003962666624122196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,512,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,256,0.00517155561182234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,256,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,256,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,128,0.004843555390834808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,128,0.0032302221904198327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,128,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,64,0.0031928889867332247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,64,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,3072,32,0.0032559999575217566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,3072,32,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,65536,0.14620888233184814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,65536,0.13121244642469618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,16384,0.05944977866278755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,65536,0.14496978123982748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,16384,0.039451556073294744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,16384,0.03721333212322659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,12288,0.029647111892700195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,16384,0.03842755489879184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,12288,0.032079110542933144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,12288,0.029458665185504492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,10240,0.028573334217071533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,10240,0.02372444503837162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,12288,0.048431111706627734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,10240,0.02528000043498145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,8192,0.024341333243581984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,8192,0.02124711043304867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,8192,0.021519111262427434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,7168,0.022299556268586054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,7168,0.018665777312384713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,7168,0.019510222805870902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,6144,0.016718222035302054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,6144,0.019744000501102872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,6144,0.017435555656750996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,5120,0.017847999930381775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,4096,0.014834667245546976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,5120,0.015592889653311836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,4096,0.012254222399658628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,4096,0.013012444807423485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,3584,0.012601777911186218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,3584,0.011339555184046427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,3584,0.012290666500727335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,3072,0.01111288865407308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,3072,0.010240000155236986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,3072,0.011376889215575324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,2560,0.008995555341243744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,2560,0.009357333183288574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,2560,0.010460444622569613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,2048,0.008359111017651027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,2048,0.007886221839321984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,2048,0.009498666558000777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,1536,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,1536,0.006697777658700943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,1536,0.007805333369308048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,1024,0.005375111268626318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,1024,0.00686488880051507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,768,0.004470222112205294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,768,0.005446222093370225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,768,0.006511999915043513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,512,0.003907555507289039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,512,0.006159111029571957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,256,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,256,0.0034755555291970572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,256,0.005818666683302985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,128,0.004480000171396467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,128,0.003167111012670729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,128,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,64,0.0030133333057165146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,64,0.0058711109062035876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2560,32,0.003040888864133093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2560,32,0.006031999985376994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,65536,0.14497244358062744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,65536,0.10295999712414211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,65536,0.1330044402016534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,16384,0.03651644455062018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,16384,0.03345955411593119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,16384,0.03451199995146857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,12288,0.028287112712860107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,12288,0.027473777532577515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,12288,0.025620443953408137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,10240,0.02286044425434536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,6144,0.0255057778623369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,10240,0.024421332610978022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,7168,0.02849688794877794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,10240,0.022574222750133936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,8192,0.020036444067955017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,8192,0.0196435550848643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,8192,0.023327999644809302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,7168,0.018054222067197163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,7168,0.017343110508388944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,7168,0.021021333005693223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,6144,0.018368888232443068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,6144,0.01605422298113505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,5120,0.013669333524174161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,5120,0.016062221593326993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,5120,0.013919111755159167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,2560,0.012784000072214337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,4096,0.013599110974205865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,4096,0.011884444289737277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,3072,1536,0.009441778063774109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,3584,0.012296888563368054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,3584,0.011311111350854238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,3072,0.009961778091059791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,3072,0.011043555206722684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,3072,0.01068800025516086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,2560,0.008912888665994009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,2560,0.008719111482302347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,2560,0.009945777555306753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,2048,0.0075128889746136135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,2048,0.0074720001882976955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,2048,0.008752889103359645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,1536,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,1536,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,1536,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,1024,0.0057884446448749965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,1024,0.004836444639497333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,1024,0.006704000135262807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,768,0.005473777651786804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,768,0.0041351111398802865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,768,0.006416000011894438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,512,0.005146666533417172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,512,0.003718222181002299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,512,0.006146666490369373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,256,0.0047306666771570844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,256,0.0033404444240861465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,256,0.005763555566469829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,128,0.004580444345871608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,128,0.003055111194650332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,128,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,64,0.0029315555261241067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,64,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,2048,32,0.0030826667530669104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,2048,32,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,65536,0.08107288678487141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,65536,0.11484977934095596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,16384,0.027298667364650305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,65536,0.09230489201015896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,16384,0.027505778604083594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,16384,0.026498667067951624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,12288,0.022612444228596155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,12288,0.021150221427281696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,10240,0.018792000081804063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,10240,0.02678311036692725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,10240,0.019091554813914828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,8192,0.015787555111779105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,8192,0.021370665894614324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,8192,0.016358221570650738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,7168,0.015608888533380298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,7168,0.01923466722170512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,7168,0.015258666541841296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,6144,0.013645333548386892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,6144,0.01684977776474423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,6144,0.013983110586802164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,5120,0.012236444486512078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,5120,0.014873777826627096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,5120,0.012824888858530255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,4096,0.010486221975750394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,4096,0.012464889221721225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,4096,0.011184000306659274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,3584,0.010145777629481422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,3584,0.011314666933483548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,3584,0.010604444477293225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,3072,0.00889955543809467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,3072,0.010279110736317104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,3072,0.009354666703277165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,2560,0.007902222375075022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,5120,0.014420444766680399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,2560,0.007747555772463481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,2560,0.008388444781303406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,2048,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,2048,0.006888000087605582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,2048,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,1536,0.00554577757914861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,1536,0.0064284445510970215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,1536,0.007096000015735626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,1024,0.00573600000805325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,1024,0.004392888810899523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,1024,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,768,0.005341333233647876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,768,0.003999111139112049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,768,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,512,0.0037120001183615792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,512,0.005011555635266834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,512,0.00608711110221015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,256,0.004634666774008009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,256,0.003313777761326896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,256,0.005709333138333426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,128,0.004452444612979889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,128,0.002971555623743269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,512,0.005074666606055366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,128,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,64,0.005719999886221356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,64,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1536,32,0.002983111060327954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1536,32,0.005720888988839255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2560,1024,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,65536,0.059824890560574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,16384,0.04662933283381992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,65536,0.08557155397203232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,65536,0.18911288844214547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,16384,0.02154133386082119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,16384,0.024632889363500807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,12288,0.0203137778573566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,12288,0.020925333102544148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,10240,0.03265155686272515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,10240,0.018391110830836825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,10240,0.018233777748213876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,8192,0.02648977769745721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,8192,0.015247111519177755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,8192,0.015599111715952555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,7168,0.013746666411558786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,7168,0.02327022287580702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,7168,0.014648000399271647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,6144,0.020212444994184706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,6144,0.012723555167516073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,6144,0.013590222431553734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,5120,0.011113777756690979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,5120,0.017416889468828838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,5120,0.012427555190192329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,4096,0.01404266721672482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,4096,0.009541333549552495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,4096,0.010590222146775989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,3584,0.012637333737479316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,3584,0.008801778157552084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,3584,0.009583999713261923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,3072,0.008232889076073965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,3072,0.011331555744012197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,3072,0.008744888835483128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,2560,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,2560,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,2560,0.008047999607192146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,2048,0.00884622252649731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,2048,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,2048,0.007432888779375289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,1536,0.005356444252861871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,1536,0.007784888976150089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,1536,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,1024,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,1024,0.004404444247484207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,1024,0.0064311110311084324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,768,0.0058453331391016645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,768,0.004019555532270008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,768,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,512,0.0035911111368073356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,512,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,256,0.004824888788991504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,256,0.003181333343187968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,256,0.005704000178310607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,128,0.004508444418509801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,128,0.0029253332565228143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,128,0.005509333478079901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,64,0.002790222151411904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,64,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,1024,32,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,1024,32,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,65536,0.15091644393073186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,65536,0.04887110988299052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,65536,0.0856319997045729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,16384,0.021772444248199463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,16384,0.043860445419947304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,16384,0.02496355606449975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,12288,0.0359191099802653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,6144,0.01573155489232805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,12288,0.018230222993426852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,12288,0.02010577751530541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,10240,0.015824889143308003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,4096,0.011763555308183035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,10240,0.018135999639829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,8192,0.025157334076033697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,8192,0.01551555593808492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,8192,0.01332355539004008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,7168,0.014524444937705994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,7168,0.01257066677014033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,6144,0.012044444680213928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,6144,0.013183111117945777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,5120,0.01667555504375034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,5120,0.010593777729405297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,5120,0.012010667059156628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,4096,0.013975110318925647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,4096,0.009275555610656738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,4096,0.009866666462686326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,3584,0.0122097780307134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,3584,0.008475555313958062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,3584,0.009240888886981541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,3072,0.01096088853147295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,3072,0.007719111111429002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,3072,0.008663111262851292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,2560,0.00981155534585317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,2560,0.0068195557428730865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,2048,3584,0.011039999624093374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,2560,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,2048,0.008610666626029545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,2048,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,2048,0.0074275558193524676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,1536,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,1536,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,1024,0.006120000034570694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,1024,0.004192000048028098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,1024,0.006501333167155583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,768,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,768,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,768,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1536,12288,0.02225333286656274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,512,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,512,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,512,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,256,0.0031386667655573953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,256,0.004744889007674323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,256,0.005647111270162795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,128,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,128,0.004419555680619346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,128,0.005490666462315454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,64,0.0027137777457634607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,64,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,768,32,0.0027822222974565295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,768,32,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,65536,0.03793333305252923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,65536,0.0758755538198683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,16384,0.023043556345833674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,65536,0.07590399848090278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,16384,0.01555022266176012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,16384,0.02332088847955068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,12288,0.018976888722843595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,12288,0.014889778362380134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,12288,0.01945244438118405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,10240,0.015944888194402058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,10240,0.013344888885815939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,10240,0.017429333594110277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,8192,0.014755555325084262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,8192,0.011780444118711682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,8192,0.01499644418557485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,7168,0.012998222476906247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,7168,0.011085333095656501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,7168,0.01367288910680347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,6144,0.01146399974822998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,6144,0.010432889064153036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,6144,0.012828444441159567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,5120,0.010117332968446944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,5120,0.009642666412724389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,5120,0.010947555303573608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,4096,0.008880889250172509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,4096,0.009675555759006077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,3584,0.008033778104517195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,3584,0.008393777741326226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,3584,0.0088764445649253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,12288,0.03755999935997857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,3072,0.007551111280918121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,3072,0.008424888882372115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,2560,0.0070062221752272705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,2560,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,2560,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,2048,0.006486222147941589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,2048,0.005507555686765247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,2048,0.00739733295308219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,1536,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,1536,0.004784000002675586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,1536,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,1024,0.005396444350481033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,1024,0.004110222061475118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,1024,0.006456888798210356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,768,0.005048888838953442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,768,0.0037946667936113146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,768,0.006227555374304454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,512,0.004761777818202972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,512,0.003478222423129611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,1024,512,0.0053093334039052325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,512,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,256,0.004405333350102107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,256,0.003112888791494899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,256,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,128,0.004283555679851108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,128,0.005487999982304043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,128,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,64,0.0027297778675953546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,512,32,0.0027466666781240036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,64,0.005482666608360078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,512,32,0.005516444643338521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,65536,0.027998222245110407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,65536,0.07627911037868924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,16384,0.02237777743074629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,16384,0.011758222348160215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,16384,0.0233297778500451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,12288,0.017996443642510306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,12288,0.010288889209429422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,12288,0.019107555349667866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,10240,0.015602666470739575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,10240,0.009581333233250512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,10240,0.017461333009931777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,8192,0.012512889173295764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,6144,0.019816888703240287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,8192,0.008728889127572378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,8192,0.014299554957283868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,7168,0.022795556320084467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,10240,0.030361778206295435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,7168,0.008417777717113495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,7168,0.013341333303186627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,6144,0.010281778044170803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,6144,0.007463999920421177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,6144,0.011692444483439127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,5120,0.008014221986134848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,5120,0.010527110762066312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,4096,0.007135111424658034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,4096,0.009379555781682333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,3584,0.00778399987353219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,3584,0.0070017774899800616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,3584,0.008958222137557136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,3072,0.007191999918884701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,3072,0.006729777902364731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,3072,0.008430221842394935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,2560,0.00666844430896971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,2560,0.006116444451941385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,2560,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,2048,0.006157333238257303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,2048,0.005400889035728242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,2048,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,1536,0.005759999983840519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,1536,0.0047253333032131195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,768,1536,0.0075351107451650835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,1536,0.006832888970772426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,1024,0.005220444252093633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,1024,0.0040693332751592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,1024,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,768,0.003714666598372989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,768,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,512,0.00462666650613149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,512,0.0034231110993358824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,512,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,256,0.004334222111437055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,256,0.003015999992688497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,256,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,128,0.0042133331298828125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,128,0.0028515555378463534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,128,0.005432888865470886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,64,0.0027093332674768236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,64,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,256,32,0.002711999954448806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,256,32,0.005470222069157495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,65536,0.07500000132454766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,65536,0.021705778108702764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,65536,0.0751146674156189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,16384,0.02211644417709774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,16384,0.010033778018421596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,16384,0.023058666123284235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,12288,0.01720711092154185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,12288,0.00889955543809467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,12288,0.018767111831241183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,10240,0.014691554837756686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,10240,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,10240,0.016032889485359192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,8192,0.01220533334546619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,8192,0.007257777783605788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,8192,0.01310755560795466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,7168,0.011143999795118967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,7168,0.007194666398896112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,7168,0.012151999606026543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,6144,0.010144888526863521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,6144,0.0064560001095136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,6144,0.01108888867828581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,5120,0.009142222503821055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,5120,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,5120,0.010311110979980892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,4096,0.008117333054542542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,4096,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,4096,0.009159111314349705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,3584,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,3584,0.006907555378145642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,3584,0.00869955536392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,3072,0.007071110937330458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,3072,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,3072,0.008406221866607666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,2560,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,2560,0.005964444329341252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,2560,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,2048,0.006057777752478917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,2048,0.005340444544951121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,2048,0.007106666763623555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,1536,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,1536,0.004646222210592694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,1536,0.00665244460105896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,1024,0.005065777649482091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,1024,0.003961777935425441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,1024,0.0061928890645504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,768,0.004835555536879433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,768,0.003608889049953885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,768,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,512,0.004559111264016894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,512,0.0032835555159383347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,512,0.005751111027267244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,256,0.004326222257481681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,4096,0.008743111458089616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,256,0.002982222164670626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,256,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,128,128,0.0041448887851503165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,128,0.0027742222365405825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,128,0.005382222019963794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,64,0.0026435556097163092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,64,0.005434666656785541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,512,3072,0.007582222421964009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,128,32,0.0026506665680143568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,128,32,0.005403555515739653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,65536,0.017715555098321702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,16384,0.008450667063395182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,65536,0.07426133420732287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,16384,0.023152889476882085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,12288,0.018290667070282828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,12288,0.007115555306275685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,10240,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,10240,0.015476443701320224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,8192,0.006784889019197888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,8192,0.01274577776590983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,7168,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,7168,0.011887999872366587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,6144,0.006405333264006509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,5120,0.006791111081838608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,6144,0.010938666760921478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,5120,0.01016711112525728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,4096,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,4096,0.009119110802809397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,3584,0.006777777853939269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,3584,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,3072,0.0064702220261096954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,3072,0.008222222328186035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,2560,0.007633777956167857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,2560,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,2048,0.007083555890454187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,2048,0.0052968888647026485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,1536,0.004594666676388847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,1536,0.006672888994216919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,1024,0.003925333420435588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,1024,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,768,0.0035706667436493766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,768,0.006000888844331105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,512,0.003261333331465721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,512,0.005763555566469829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,256,0.00291377781993813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,256,0.005538666827811136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,128,0.002732444347606765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,128,0.005400889035728242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,64,0.0026106666773557663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,64,0.005443555613358815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,64,32,0.0026106666773557663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,64,32,0.005399111244413588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,65536,0.07518844472037421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,16384,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,16384,0.023092443744341534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,65536,0.01642933322323693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,12288,0.0070773329999711775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,10240,0.006739555547634761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,12288,0.017823111679818895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,8192,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,10240,0.014913777510325113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,8192,0.01276000009642707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,7168,0.006393777827421824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,7168,0.011966221862369113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,6144,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,6144,0.011007111105653973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,5120,0.010113777385817634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,5120,0.006698666761318843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,4096,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,3584,0.006687110910813014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,4096,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,3584,0.008538666698667739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,3072,0.0063573333124319715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,3072,0.008147555920812819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,2560,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,2560,0.007568000091446771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,2048,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,2048,0.007113777928882175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,1536,0.00461422238085005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,1536,0.006559111177921295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,1024,0.003943110919660992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,1024,0.006161777923504512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,768,0.003559111307064692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,768,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,512,0.0032826666202810076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,512,0.005710222240951326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,256,0.002924444360865487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,256,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,128,0.0026933333526055017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,128,0.00534400012758043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,64,0.0025928889711697898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,64,0.005366222312053044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,160,32,32,0.002573333266708586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,160,32,32,0.00535111129283905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,16384,0.7296000056796603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,65536,0.0752355522579617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,16384,0.38632623354593915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,16384,0.40255289607577854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,12288,0.5584168963962132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,12288,0.3236364523569743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,10240,0.2325191232893202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,10240,0.4663075341118707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,12288,0.31025510364108616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,7168,0.011324444578753578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,8192,0.3733679983350966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,5120,0.009219555391205682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,4096,0.008173333273993598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,8192,0.22000622749328613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,10240,0.25576533211602104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,7168,0.3266399966345893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,8192,0.21019555462731254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,7168,0.16758667098151314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,6144,0.28413955370585126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,7168,0.17674222257402208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,6144,0.16175022390153673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,5120,0.23761243290371367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,5120,0.12545866436428493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,6144,0.16848444938659668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,4096,0.10421066813998753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,5120,0.1303208933936225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,4096,0.19374399714999727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,4096,0.10328266355726455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,3584,0.09311555491553412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,3584,0.17720088693830702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,3072,0.15398222870296902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,160,256,768,0.004920888692140579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,3584,0.09141688876681858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,2560,0.12909421655866835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,2560,0.07072977887259589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,3072,0.08498933580186631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,2048,0.10166755649778579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,2048,0.0647688905398051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,1536,0.07921599679523043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,2560,0.07656088802549574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,1024,0.03495111068089803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,1536,0.04814844330151876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,1536,0.046112888389163546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,2048,0.05740444527732002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,1024,0.05604800250795153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,768,0.04429511229197184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,1024,0.03966755668322245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,768,0.029224889145957098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,512,0.033055110110176936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,512,0.028303109937243994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,256,0.019466666711701285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,512,0.023786667320463393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,768,0.03256266646915012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,256,0.02330933345688714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,128,0.013849777479966482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,128,0.011710222396585675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,64,0.00920800036854214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,128,0.02117777698569828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,64,0.02567911148071289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,65536,32,0.008993777963850234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,65536,32,0.028185778194003638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,16384,0.19280621740553114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,16384,0.09830577505959405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,16384,0.10705244541168213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,12288,0.13905155658721924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,12288,0.0823777781592475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,12288,0.0820853312810262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,65536,0.7452364497714572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,10240,0.12447378370496963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,10240,0.06458222203784518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,10240,0.07039644320805867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,8192,0.10942311419381036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,8192,0.060772445466783315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,8192,0.055698666307661265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,7168,0.09714666340086196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,6144,0.04230577747027079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,7168,0.04964800013436211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,6144,0.07647644148932563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,65536,0.4228257868025038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,5120,0.03680266605483161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,6144,0.043567998541726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,5120,0.0628124475479126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,4096,0.05011644297175937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,4096,0.031160887744691636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,4096,0.03331022130118476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,5120,0.0388648874229855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,3584,0.028024888700909082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,3584,0.04459733433193631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,3584,0.02962488929430644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,3072,0.039107554488711886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,3072,0.026085333691702947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,3072,0.025411556164423626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,2560,0.03355822298261855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,2560,0.023208000593715247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,2048,0.027824888626734417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,2048,0.01926488843229082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,1536,0.02247911029391819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,1536,0.016139555308553908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,2048,0.01976977785428365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,1536,0.015810666812790763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,1024,0.013068444199032254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,1024,0.013497778111033969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,1024,0.01646488904953003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,768,0.011607999602953592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,768,0.013935110635227628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,768,0.01290933373901579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,512,0.01035911093155543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,512,0.010455999937322406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,512,0.011337777806652917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,256,0.006674666785531574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,256,0.009263111485375298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,256,0.00835111074977451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,128,0.004517333375083076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,128,0.00742222203148736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,128,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,64,0.004303110970391167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,64,0.008798222574922774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,16384,32,0.004763555609517627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,16384,32,0.009860444400045607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,65536,0.5523262023925781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,16384,0.0787351131439209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,65536,0.3432568973965115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,16384,0.13596533404456243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,12288,0.06117510795593262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,16384,0.0910124447610643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,12288,0.10393689076105754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,12288,0.06333244509167142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,10240,0.052148444785012134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,10240,0.08816978004243638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,8192,0.04295555419392056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,3072,0.08166399929258558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,10240,0.05449600021044413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,8192,0.07053066624535455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,7168,0.0624542236328125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,8192,0.04381688767009311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,7168,0.03884177737765842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,6144,0.05531555414199829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,65536,256,0.018556444181336295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,7168,0.039008888933393694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,6144,0.03436444534195794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,5120,0.04680000080002678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,6144,0.03484444485770331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,5120,0.02978044417169359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,5120,0.030416889323128596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,4096,0.024847999215126038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,4096,0.038189331690470375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,4096,0.025429333249727886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,3584,0.03418933351834615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,3584,0.022981334063741896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,3072,0.030386666456858318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,3072,0.02057422200838725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,2560,0.01812533371978336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,2560,0.025974222355418738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,2560,0.01831288966867659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,2048,0.021623111433453027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,2048,0.01585155559910668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,2048,0.015750222735934787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,1536,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,1536,0.017317333155208163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,1536,0.01353422221210268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,1024,0.011988444460762871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,1024,0.01129244433508979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,1024,0.01185955521133211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,768,0.010211555494202508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,768,0.009931555224789513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,768,0.010936000280910067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,512,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,512,0.007676444119877285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,512,0.009682666924264696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,256,0.005505777895450592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,256,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,256,0.007511999871995714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,128,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,128,0.004108444270160464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,64,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,128,0.0070462218589252895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,64,0.007807999849319458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,12288,32,0.008594666918118795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,12288,32,0.004291555533806483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,65536,0.47966935899522567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,7168,0.04797333478927612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,65536,0.3625226550632053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,16384,0.06810666455162896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,16384,0.13260444005330405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,16384,2560,0.02234222160445319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,12288,0.08874222305085923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,16384,0.07785955402586195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,12288,0.060630222161610924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,10240,0.045639998382992215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,10240,0.07463377714157104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,65536,0.3210151195526123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,10240,0.05096177922354805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,8192,0.06957511107126872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,8192,0.04166400101449754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,7168,0.03433777888615926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,7168,0.05300355619854397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,7168,0.03702577617433336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,6144,0.03025777803526984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,6144,0.047093331813812256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,6144,0.03329066766632928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,5120,0.026211554805437725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,5120,0.039742221434911094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,5120,0.028446220689349707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,4096,0.021948445174429152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,4096,0.03264888789918687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,4096,0.02362222141689724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,3584,0.020405333903100755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,3584,0.029209779368506536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,3584,0.02161066730817159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,3072,0.018237334158685472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,3072,0.026016889346970454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,3072,0.01944977707333035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,2560,0.016474665866957772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,2560,0.022320000661744013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,2560,0.01700444519519806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,2048,0.014361777239375643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,2048,0.01852888862291972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,2048,0.015030221806632148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,65536,0.2878897719913059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,1536,0.012488889197508493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,1536,0.016547555724779766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,1536,0.013101333545313941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,1024,0.010506666368908353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,1024,0.011349333657158745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,768,0.009480888644854227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,768,0.008921778036488427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,768,0.010443555812040964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,512,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,512,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,512,0.0084906667470932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,256,0.005697777701748743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,256,0.004842666702138053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,256,0.0074755557709270054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,128,0.005163555757866966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,128,0.003923555629120933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,128,0.007056000332037608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,64,0.003714666598372989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,64,0.007799111306667328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,10240,32,0.0041493334703975255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,10240,32,0.008415111237102086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,65536,0.204438222779168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,65536,0.400948445002238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,3584,0.02304800020323859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,12288,3072,0.020489777127901714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,16384,0.10372977786593968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,16384,0.0621582203441196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,16384,0.057096891933017306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,65536,0.22205866707695854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,12288,0.07133244143591987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,12288,0.04511555698182848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,10240,0.06004889143837822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,10240,0.036883556180530123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,8192,0.030552887254291113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,8192,0.0485208895471361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,10240,0.03896444373660617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,7168,0.04303110970391167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,7168,0.027464000715149775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,8192,0.03219555483924018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,6144,0.03841777642567953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,7168,0.028811554114023846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,6144,0.024481778343518574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,5120,0.03243288728925917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,6144,0.025897777742809717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,5120,0.021341333786646526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,4096,0.026871111657884385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,5120,0.02277155551645491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,4096,0.01847022275129954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,3584,0.023948444260491267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,4096,0.01907111042075687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,3584,0.017107556263605755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,3584,0.017639999588330586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,3072,0.021427555216683283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,3072,0.015909334023793537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,2560,0.014019555515713163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,2560,0.01832177738348643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,2560,0.014299554957283868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,2048,0.012583110895421771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,2048,0.015225778023401896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,2048,0.012778666284349231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,1536,0.011048888994587792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,1536,0.011864888999197217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,1536,0.011340444286664328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,1024,0.009377777576446533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,1024,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,1024,0.009900444083743626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,768,0.007712889048788283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,768,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,768,0.009134222235944536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,512,0.007287110719415877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,512,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,512,0.007174222005738153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,256,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,256,0.004063111212518481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,256,0.006539555473460092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,128,0.0036275556517971885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,128,0.004942222187916438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,128,0.006337777607970768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,64,0.0034248888906505373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,64,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,8192,32,0.0037617778612507712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,8192,32,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,65536,0.24825334548950195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,65536,0.34882399770948624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,12288,0.05339822173118591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,65536,0.1774533324771457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,16384,0.09184710847006904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,65536,0.2205653323067559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,8192,0.037520888778898455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,16384,0.053935110569000244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,12288,0.07243021991517809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,12288,0.041871110598246254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,12288,0.043883555465274386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,10240,0.0535733335547977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,10240,0.03696977761056688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,8192,0.027951998843087092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,8192,0.04321066538492838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,8192,0.030068443881140813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,7168,0.03880177934964498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,7168,0.025340444511837427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,7168,0.027179555760489568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,6144,0.034690668185551964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,6144,0.024358222881952923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,5120,0.01962844365172916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,5120,0.02921688887808058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,5120,0.021216000119845074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,4096,0.017039111918873258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,4096,0.023896889554129705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,4096,0.018064889642927382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,3584,0.015815110670195687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,3584,0.021328000558747187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,3584,0.016740444633695815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,3072,0.014509333504570855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,3072,0.018924444913864136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,3072,0.015141333142916361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,2560,0.013187555803192986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,2560,0.016202666693263583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,2560,0.013521778086821238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,2048,0.013331555657916598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,2048,0.012302222351233164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,2048,0.011940444509188334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,10240,1024,0.010346666806273991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,1536,0.010676444404655032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,1536,0.011069333387745751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,1536,0.010471999645233154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,1024,0.008239111138714684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,1024,0.009736888938479954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,1024,0.009332444104883406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,768,0.00828888929552502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,768,0.007871111234029135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,768,0.007039111521508958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,512,0.006513777706358168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,512,0.005858666780922148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,512,0.007009777757856581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,256,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,256,0.003834666477309333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,256,0.006533333410819371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,128,0.003563555578390757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,128,0.00498488876554701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,64,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,128,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,64,0.006679111056857639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,7168,32,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,7168,32,0.007069333559936947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,65536,0.28845421473185223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,16384,0.05687999725341797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,65536,0.1779564486609565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,65536,0.19837778144412568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,16384,0.07820889022615221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,12288,0.054679112301932446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,16384,0.05286488930384318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,12288,0.03956800036960178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,12288,0.04068444503678216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,10240,0.04639110962549845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,10240,0.03386933273739285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,8192,0.03738755649990506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,8192,0.027941332923041448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,10240,0.03459111187193129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,8192,0.029030223687489826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,7168,0.03331289026472304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,7168,0.02612000041537815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,6144,0.02221244408024682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,6144,0.029934220843844946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,6144,0.02367466688156128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,5120,0.019472888774342008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,5120,0.025370667378107708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,5120,0.020597333709398907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,4096,0.01681600014368693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,4096,0.02103288968404134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,4096,0.01771288944615258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,3584,0.018800000349680584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,3584,0.01605688863330417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,12288,0.043290668063693576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,3072,0.01676266723208957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,3072,0.014523555835088095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,2560,0.012977778083748288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,2560,0.014064888159434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,2560,0.013295110729005603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,2048,0.011476444701353708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,2048,0.011728888584507836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,2048,0.012010667059156628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,8192,3072,0.015472888946533203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,1536,0.00959822204377916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,1536,0.010797333386209277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,16384,0.050256000624762646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,1024,0.0075457774930530125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,10240,0.033702221181657575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,1024,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,7168,6144,0.022403554783927068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,768,0.006491555521885554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,768,0.007371555599901412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,512,0.004867555366622077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,512,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,256,0.004984000076850255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,256,0.003671110918124517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,256,0.006575111299753189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,7168,0.025223111112912495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,128,0.003310222178697586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,3584,0.015632000234391954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,128,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,64,0.003132444495956103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,64,0.006794666664467917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,6144,32,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,6144,32,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,3072,0.014373333917723762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,65536,0.25265688366360134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,65536,0.13303466637929282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,16384,0.04299022091759575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,1536,0.010244444840484196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,16384,0.06721333662668864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,16384,0.052341332038243614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,12288,0.0543066660563151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,10240,0.026142223013771906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,12288,0.03968711031807794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,65536,0.1917724476920234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,10240,0.04672799838913811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,10240,0.03401422169473436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,8192,0.032207111517588295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,1024,0.00868266655339135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,8192,0.024105777343114216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,8192,0.027756444282001917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,7168,0.01977066695690155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,7168,0.02905511193805271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,7168,0.02589688864019182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,6144,0.02606933315594991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,6144,0.017878222796652053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,6144,0.02349333299530877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,5120,0.022138666775491502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,5120,0.015984000431166757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,5120,0.020056888461112976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,4096,0.013960889644092984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,4096,0.018087999688254464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,4096,0.01698844465944502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,3584,0.016196444630622864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,3584,0.015690666106012132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,3072,0.01418133411142561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,3072,0.012159999873903064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,3072,0.01462933255566491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,2560,0.011195555329322815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,2560,0.012048888537618848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,2560,0.012998222476906247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,2048,0.010272000398900773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,2048,0.011746666497654386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,1536,0.009130666653315226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,1536,0.008536888493431939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,1536,0.010753778119881948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,1024,0.00816000004609426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,1024,0.006853333363930385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,1024,0.008257778154479133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,768,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,768,0.006027555714050929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,768,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,512,0.005512889060709212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,512,0.004408888932731417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,512,0.006836444553401735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,256,0.004962666581074397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,256,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,256,0.006481777876615524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,128,0.003236444460021125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,128,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,64,0.003048000029391713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,64,0.006523555351628198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,5120,32,0.0031679999083280563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,5120,32,0.006789333290523953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,65536,0.11149333582984076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,65536,0.1947208907869127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,768,0.007120889094140794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,16384,0.03277066681120131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,16384,0.05806844764285617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,65536,0.11491644382476807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,16384,0.03379822108480666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,12288,0.04400888747639126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,12288,0.025617778301239014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,12288,0.02757511039574941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,10240,0.022036444809701707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,10240,0.0332640012105306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,10240,0.0240320000383589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,8192,0.018756444255510967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,8192,0.026799110902680293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,512,0.005410666680998272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,8192,0.020231111182106864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,7168,0.024031110935741003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,7168,0.018556444181336295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,6144,0.015671999918089975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,6144,0.021474666065639917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,6144,0.0170151111152437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,5120,0.014019555515713163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,5120,0.01850311126973894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,5120,0.015255111787054272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,4096,0.01236533373594284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,4096,0.01514844430817498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,4096,0.013188444077968597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,6144,128,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,3584,0.01353422221210268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,3584,0.012387555506494312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,3072,0.010954666468832227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,3072,0.012010667059156628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,3072,0.011669333610269757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,2560,0.010199999643696679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,2560,0.01054044481780794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,2560,0.011053333679835001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,2048,0.009186666872766282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,2048,0.009013333254390294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,2048,0.009948444035318162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,1536,0.008265777594513362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,1536,0.007719111111429002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,1536,0.008924444516499838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,1024,0.006046222315894232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,1024,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,1024,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,768,0.005399999933110342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,768,0.005138666679461797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,768,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,512,0.005166222237878376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,512,0.0038417776425679526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,512,0.006063999815119638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,256,0.0046942221621672315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,256,0.003458666718668408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,256,0.005797333187527127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,128,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,128,0.004538666870858935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,128,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,64,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,64,0.003063999944263034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,4096,32,0.0034044444974925783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,4096,32,0.006166222194830577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,65536,0.1708764500088162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,65536,0.10168000062306722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,16384,0.029784000582165186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,65536,0.1117395559946696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,16384,0.05113155643145243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,12288,0.023413333627912734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,16384,0.03113244308365716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,12288,0.03516622384389242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,12288,0.0255724440018336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,10240,0.029512001408471003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,10240,0.020224000016848247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,12288,0.03033866816096836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,10240,0.022413333257039387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,8192,0.01732711162832048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,8192,0.023987554841571387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,8192,0.018900444110234577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,7168,0.016045333610640634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,7168,0.021807110971874658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,7168,0.017417778571446735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,6144,0.01943644384543101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,6144,0.01587377819750044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,5120,0.01667644414636824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,5120,0.014478221535682678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,2048,0.010125333236323463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,4096,0.013667555318938361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,4096,0.012470222181744046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,3584,0.013062222136391534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,3584,0.011198221809334226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,3584,0.012265777422322167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,3584,0.01184977756606208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,3072,0.010495111346244812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,5120,128,0.004774222357405556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,3072,0.01112711098459032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,3072,0.011222222612963783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,2560,0.00978311151266098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,2560,0.010148444109492833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,2560,0.010447999669445885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,2048,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,2048,0.008842666943868002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,2048,0.00958933350112703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,1536,0.007831110722488826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,1536,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,1536,0.007835555407736037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,1024,0.0058248887459437055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,1024,0.006515555497672822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,768,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,768,0.004902222090297275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,768,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,512,0.003990222182538775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,512,0.005091555416584015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,512,0.006031999985376994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,256,0.004663111021121343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,256,0.0034355556385384668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,256,0.005745777653323279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,128,0.004487110922733943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,128,0.003305777700410949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,128,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,64,0.003141333245568805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,64,0.005798222290145026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3584,32,0.0032382222513357797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3584,32,0.006024888820118374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,65536,0.09359644518958198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,65536,0.14703199598524305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,65536,0.10686577690972222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,16384,0.027181333965725366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,16384,0.04464711083306206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,16384,0.028943998946083918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,12288,0.021212443709373474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,12288,0.03478488988346524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,7168,0.017103110750516255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,12288,0.02369333306948344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,10240,0.030307554536395605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,10240,0.01887999971707662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,10240,0.020908445119857788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,8192,0.016335111525323655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,8192,0.02179644505182902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,8192,0.01763911048571269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,7168,0.014813333749771118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,7168,0.019935111204783123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,4096,3584,0.011678222152921887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,7168,0.016227554943826463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,6144,0.013911111487282647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,6144,0.017838221457269456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,6144,0.015092444088723926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,5120,0.016584888100624084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,5120,0.0125608891248703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,5120,0.013712888790501488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,4096,0.01348088930050532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,4096,0.011939555406570435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,3584,0.01074577785200543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,3584,0.01202222208182017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,3584,0.011363555159833698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,3072,0.009902222288979424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,3072,0.011031111081441244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,3072,0.0107360002067354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,2560,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,2560,0.00944800012641483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,2560,0.010106667048401302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,2048,0.008379555410808986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,2048,0.008192000289758047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,2048,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,1536,0.007184888753626082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,1536,0.006727111008432176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,1536,0.00721066693464915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,1024,0.005675555517276128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,1024,0.0064088888466358185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,1024,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,768,0.004545777622196409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,768,0.00535022219022115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,768,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,512,0.004022222426202563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,512,0.005032888717121548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,512,0.006027555714050929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,256,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,256,0.0046942221621672315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,256,0.005705777969625261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,128,0.004481777962711123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,128,0.0032302221904198327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,64,0.0029599999801980127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,128,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,64,0.005753777921199799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,3072,32,0.003140444349911478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,3072,32,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,65536,0.12544000148773193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,65536,0.08811289072036743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,65536,0.08789244625303481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,16384,0.02624622152911292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,16384,0.038400888442993164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,16384,0.027245332797368366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,12288,0.030756443738937378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,12288,0.02195466723706987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,10240,0.024030221833123103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,10240,0.01830933325820499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,10240,0.01957244508796268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,8192,0.015575110912322998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,8192,0.019910222954220243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,8192,0.016728000508414376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,7168,0.01428533262676663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,7168,0.017829333742459614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,7168,0.015493333339691162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,6144,0.013201778133710226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,6144,0.015839111473825242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,6144,0.014344000154071383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,5120,0.011944889194435544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,5120,0.013619555367363824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,5120,0.01312888910373052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,4096,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,4096,0.01144177797767851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,4096,0.011536888778209686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,3584,0.010330666270520952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,3584,0.01053422192732493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,3584,0.010942222343550788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,3072,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,3072,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,3072,0.010336000058386061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,2560,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,2560,0.008976889153321585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,2560,0.009550222092204625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,2048,0.007532444265153673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,2048,0.007486222518814935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,2048,0.008370666868156856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,1536,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,1536,0.006548444430033366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,1536,0.007118222614129384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,1024,0.005642666584915585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,1024,0.004948444250557157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,1024,0.0064311110311084324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,768,0.0053093334039052325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,768,0.004365333252482944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,768,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,5120,0.013203555511103736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,512,0.003735110991530948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,512,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,6144,0.014692443940374585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,256,0.0033982222278912864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,256,0.0057546666098965546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,128,0.004520888957712385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,128,0.003072000005178981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,128,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,64,0.0028986665937635633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,4096,0.011872889267073737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,64,0.005767111149099138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2560,32,0.0030213333666324615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2560,32,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,65536,0.10402222474416097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,65536,0.08408088816536798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3584,1024,0.006108444597986009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,16384,0.032474666833877563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,16384,0.024375110864639282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,16384,0.025798221429189045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,12288,0.0188746667570538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,12288,0.02511022157139248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,12288,0.020650666620996263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,10240,0.022080888350804646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,10240,0.018571555614471436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,8192,0.018767999278174508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,8192,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,8192,0.015896000795894198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,7168,0.013374221821626028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,7168,0.016879111528396606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,7168,0.014731556177139282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,6144,0.015474667151769003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,6144,0.012372444073359171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,6144,0.013752000199423896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,5120,0.011312888728247749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,5120,0.014504000544548035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,5120,0.01258666647805108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,4096,0.010132444401582083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,4096,0.011394666300879585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,4096,0.011136888629860349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,3584,0.009696889254781935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,3584,0.010363555616802638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,3584,0.010600888894663917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,3072,0.009020444419648912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,3072,0.009581333233250512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,3072,0.009850666754775578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,2560,0.007823111282454597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,2560,0.007824000385072496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,2560,0.008854221966531541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,2048,0.007068444457319047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,2048,0.007352888584136963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,1536,0.005979555762476391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,1536,0.005998222364319696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,1536,0.006779555645253923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,1024,0.005462222215202119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,1024,0.004632888982693355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,1024,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,768,0.005120888766315249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,768,0.0040702223777771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,768,0.006117333140638139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,512,0.004898666507667965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,512,0.00360355567600992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,512,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,256,0.004496888981925117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,256,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,256,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,128,0.004266666869322459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,128,0.002962666667169995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,128,0.005477333234416113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,64,0.0028071111689011254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,64,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,2048,32,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,2048,32,0.005677333308590784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,65536,0.07750933037863837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,65536,0.07985688580407037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,65536,0.08158844709396362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,3072,4096,0.011178666518794166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,16384,0.025133333272404138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,16384,0.024338665935728285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,12288,0.018408000469207764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,12288,0.0206977774699529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,12288,0.020184889435768127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,12288,0.02070755594306522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,10240,0.018227555685573153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,10240,0.017687110437287223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,8192,0.014313777287801107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,8192,0.015376889043384127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,8192,0.015533333023389181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,7168,0.013162666724787818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,7168,0.014244443840450711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,512,0.005031110925806893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,7168,0.01460533340771993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,6144,0.012127999630239276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,6144,0.012535110943847232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,6144,0.013253333667914072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,5120,0.011152000062995486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,5120,0.011232888533009423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,5120,0.0123848890264829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,4096,0.009962666365835402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,4096,0.00960622231165568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,4096,0.010820444259378644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,3584,0.009217778013812171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,3584,0.008908444808589088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,3584,0.010263111028406356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,3072,0.008073777788215214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,3072,0.008182221816645728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,3072,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,2560,0.00723555518521203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,2560,0.00777688870827357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,2560,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,2048,0.006525333142942852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,2048,0.006890666567616993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,2048,0.007273777491516537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,1536,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,1536,0.005523555394675996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,1536,0.006836444553401735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,1024,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2560,256,0.004632888982693355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,1024,0.004512888689835866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,1024,0.006396444307433234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,768,0.004068444586462445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,768,0.006181333214044571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,512,0.004826666580306159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,512,0.003629333443111844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,512,0.005964444329341252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,256,0.004486222234037188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,256,0.0032257777121331957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,256,0.005688000056478713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,128,0.0029608888758553397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,128,0.005516444643338521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,64,0.0028008888992998335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,64,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1536,32,0.0028595555987623003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1536,32,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,65536,0.07338755660586886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,65536,0.05667733483844333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,65536,0.0793386697769165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,65536,0.0748986668056912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,16384,0.018964444597562153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,16384,0.023183999790085688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,12288,0.017497777938842773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,10240,0.016662221815851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,12288,0.015230221880806817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,12288,0.018906666172875296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,10240,0.015446222490734525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,10240,0.013719999955760108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,10240,0.017159110969967313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,8192,0.013404444687896304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,8192,0.011853333148691388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,8192,0.014652444256676568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,7168,0.011542222566074796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,7168,0.013831111292044321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,6144,0.011405333048767514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,6144,0.012454222473833295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,6144,0.012858666479587555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,5120,0.010231999887360467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,5120,0.009917333722114563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,2048,2048,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,5120,0.011794666449228922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,4096,0.009404444032245213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,4096,0.0096133334769143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,3584,0.00794933322403166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,3584,0.008993777963850234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,3584,0.0088755554623074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,16384,0.023855111665195886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,3072,0.008361777497662438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,3072,0.008536888493431939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,2560,0.0068693334857622785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,2560,0.00685777763525645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,2560,0.007743111087216272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,2048,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,2048,0.005647111270162795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,2048,0.007175111108356052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,1536,0.005780444376998478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,1536,0.004958222309748332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,1536,0.006765333314736684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,10240,0.016228444046444364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,1024,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,1024,0.004168888760937585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,1024,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,768,0.003871110992299186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,768,0.006129777679840724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,512,0.004815111143721474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,512,0.0035333335399627686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,512,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,256,0.004456000195609199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,256,0.0031911111954185697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,256,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,128,0.004228444563017951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,128,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,128,0.005454222361246745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,64,0.002760888801680671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,64,0.00544088871942626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,1024,32,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,1024,32,0.005517333332035277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,65536,0.07305155860053168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,65536,0.045087112320793994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,65536,0.06933866606818305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,16384,0.02201866606871287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,16384,0.015842666228612263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,16384,0.02252977755334642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,12288,0.017312889297803242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,12288,0.013780444032616086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,12288,0.018576888574494254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,10240,0.015271110667122735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,10240,0.012777778009573618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,768,0.005012444323963589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,10240,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,8192,0.013242666920026144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,8192,0.011149333582984077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,8192,0.014569777581426831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,7168,0.01221244451072481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,7168,0.010584000084135266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,7168,0.01369955556260215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,6144,0.009886222581068674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,6144,0.012402666939629449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,5120,0.009314667019579146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,5120,0.009703999592198266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,5120,0.011114666859308878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1536,128,0.004297777596447202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,4096,0.008257778154479133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,4096,0.0088746663596895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,4096,0.009011555876996782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,3584,0.007699555820888943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,3584,0.008268444074524773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,3584,0.008635555704434713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,3072,0.007378666765160031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,3072,0.007175111108356052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,3072,0.008112000094519721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,2560,0.006790222393141852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,2560,0.006169777777459886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,2560,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,2048,0.005477333234416113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,2048,0.007156444092591603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,1536,0.005734222216738596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,1536,0.0047884442740016514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,1536,0.006714666469229593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,1024,0.00517511119445165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,1024,0.004104888687531154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,1024,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,768,0.0048942222363419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,768,0.0037439999481042228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,768,0.006066666709052191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,16384,0.022364444202846948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,512,0.0034684443639384378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,512,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,7168,0.012451555165979596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,256,0.003044444446762403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,256,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,128,0.004260444392760594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,128,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,128,0.005456888841258154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,4096,0.008591999610265097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,64,0.0026551110463009942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,64,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,768,32,0.0027111110587914786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,768,32,0.005526222288608551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,65536,0.03682933251063029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,65536,0.07315911187065972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,65536,0.06974222262700398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,16384,0.02187999917401208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,3072,0.00720177756415473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,16384,0.013125333521101209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,16384,0.022286223040686712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,12288,0.01146399974822998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,12288,0.01717866626050737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,12288,0.018333334061834548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,10240,0.011049778097205691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,10240,0.0166311115026474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,8192,0.012647111382749347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,8192,0.00942311104800966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,8192,0.01428711083200243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,7168,0.011885333392355176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,7168,0.009076444639099969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,7168,0.013205333716339536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,6144,0.010188444621033138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,6144,0.010022222167915767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,6144,0.011713777979214987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,5120,0.009186666872766282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,5120,0.009175111022260454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,5120,0.010332444475756751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,4096,0.008126222425036961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,4096,0.007791111038790808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,4096,0.008821333448092142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,3584,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,3584,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,3584,0.008459555606047312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,3072,0.007099555598364935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,3072,0.006755555669466655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,3072,0.00794222205877304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,1024,768,0.004956444518433677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,2560,0.006130666782458623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,2560,0.0075635554061995606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,2048,0.005478222337034013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,2048,0.007039111521508958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,1536,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,1536,0.004751111070315043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,1536,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,1024,0.005128889034191768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,1024,0.004032888760169347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,1024,0.006216889040337668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,768,0.004872888740566042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,768,0.0037324445115195382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,768,0.005994666781690385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,512,0.004615111069546806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,512,0.003370666669474708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,512,0.0057555557125144536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,256,0.00433955548538102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,256,0.0030248889492617715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,256,0.005512889060709212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,128,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,128,0.0028284444577164124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,128,0.005387555393907759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,64,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,64,0.005412444472312927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,512,32,0.00300799993177255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,512,32,0.005423999908897612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,65536,0.07224088907241821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,65536,0.025514667232831318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,65536,0.0691431098514133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,16384,0.021506667137145996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,16384,0.010665777656767102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,16384,0.02218666672706604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,6144,0.010592888626787396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,12288,0.009568000005351173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,12288,0.018208889497650992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,2048,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,10240,0.008973333570692275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,10240,0.01591199967596266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,8192,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,8192,0.008259555531872643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,8192,0.013095999757448832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,7168,0.011060444845093621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,7168,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,7168,0.012096000214417776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,6144,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,6144,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,512,0.0046844445168972015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,6144,0.010803555448849997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,5120,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,5120,0.010112000008424124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,4096,0.008037333687146505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,4096,0.006532444308201472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,4096,0.008923555413881937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,3584,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,3584,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,3584,0.008473777936564552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,768,256,0.004454222404294544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,3072,0.007108444141017065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,3072,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,3072,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,2560,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,2560,0.006048000107208888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,2560,0.00758844448460473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,2048,0.006051555689838197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,2048,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,2048,0.007058666812049017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,1536,0.0046977777447965415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,1536,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,1536,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,1024,0.004015111260943943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,1024,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,768,0.0047884442740016514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,768,0.003669333126809862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,768,0.005980444451173146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,512,0.004583999928500918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,512,0.0033582221302721235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,512,0.005759999983840519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,256,0.004313777718279096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,256,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,256,0.005525333185990651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,128,0.004204444587230682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,128,0.002790222151411904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,128,0.005390222287840313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,64,0.0026497778793176017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,64,0.005416888743638992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,256,32,0.0026666666898462507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,256,32,0.005397333453098933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,65536,0.07246933380762736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,65536,0.020131554868486192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,10240,0.015215999550289579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,65536,0.06895555390252008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,16384,0.021451556020312842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,16384,0.008909333083364699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,16384,0.022129777404997084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,12288,0.016399999459584553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,12288,0.00758844448460473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,12288,0.017493334081437852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,10240,0.014053333136770459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,10240,0.007266666326257918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,10240,0.014680888917711047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,2560,0.006703111032644908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,8192,0.006706666615274217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,8192,0.0123831108212471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,512,2048,0.006109333286682765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,7168,0.00657155571712388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,7168,0.011427555647161273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,6144,0.009961778091059791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,6144,0.00638133328821924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,6144,0.01055555542310079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,5120,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,5120,0.009824000298976898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,4096,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,4096,0.006401777681377199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,4096,0.008852444589138031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,3584,0.007535999847782983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,3584,0.0069297779765393995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,3584,0.008366222182909647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,3072,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,3072,0.00654755574133661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,3072,0.007905777957704332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,2560,0.006528000036875407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,2560,0.006032889087994893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,2560,0.007490666376219855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,2048,0.006049777898523543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,2048,0.005373333477311664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,2048,0.007038222418891058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,1536,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,1536,0.0046693334976832075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,1536,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,1024,0.00498577786816491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,1024,0.004020444634887907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,1024,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,768,0.0047591109242704184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,768,0.003663111064169142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,768,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,512,0.004505777938498391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,512,0.0033324443631701996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,512,0.00572533326016532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,256,0.004271999829345279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,256,0.002973333415057924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,256,0.005491555564933353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,128,0.004081777814361784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,128,0.0027795556104845474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,128,0.0054551110499435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,64,0.0026177778426143858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,64,0.005393777870469623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,128,32,0.0026355555488003623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,128,32,0.005387555393907759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,65536,0.015985776980717976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,65536,0.06942044364081489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,16384,0.00741155528359943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,16384,0.022117333279715642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,12288,0.006559999866618051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,12288,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,10240,0.00666222224632899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,10240,0.014244443840450711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,8192,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,8192,0.012215110990736218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,7168,0.0064071110553211635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,7168,0.011472889118724398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,6144,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,6144,0.010595555934641095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,5120,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,5120,0.00979644474056032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,4096,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,4096,0.008813333180215623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,3584,0.006711999989218182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,3584,0.008411555654472774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,3072,0.0063768890168931745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,3072,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,2560,0.006028444402747684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,2560,0.007495111061467065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,2048,0.005333333379692501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,2048,0.0070222218831380205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,1536,0.004680888934267892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,1536,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,1024,0.003989333493842019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,12288,0.01681422193845113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,1024,0.006118222243256039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,768,0.003628444340493944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,768,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,512,0.003293333368168937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,512,0.005711110929648082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,256,0.0029680000411139596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,256,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,128,0.002769777758253945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,128,0.0053546664615472155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,64,0.002624888800912433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,64,0.005372444374693765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,64,32,0.002616000051299731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,64,32,0.005396444350481033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,65536,0.014577777849303352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,16384,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,65536,0.06926844517389934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,16384,0.021729777256647747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,12288,0.006365333166387346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,10240,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,12288,0.016380444169044495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,8192,0.006555555595291986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,10240,0.013950222068362765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,8192,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,7168,0.006342222293217977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,7168,0.011307555768224927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,6144,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,6144,0.010464000205198923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,5120,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,5120,0.00964355551534229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,4096,0.006077333456940121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,4096,0.008610666626029545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,3584,0.006515555497672822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,3584,0.008242666721343994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,3072,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,3072,0.007868444754017724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,2560,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,2560,0.0074720001882976955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,2048,0.005255110975768831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,2048,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,1536,0.004619555754794015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,1536,0.006559111177921295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,1024,0.003943110919660992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,1024,0.006072888771692912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,768,0.0035964445107513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,768,0.0058542220956749385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,512,0.0032888888898822996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,512,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,256,0.0029253332565228143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,256,0.005443555613358815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,128,0.0027066665805048416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,128,0.0053137776752312975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,64,0.0025946667624844448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,64,0.005340444544951121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,128,32,32,0.002604444407754474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,128,32,32,0.005327111317051782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,16384,0.3820826742384169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,16384,0.7221848699781629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,10240,0.014319111903508505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,5120,0.00907022257645925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,12288,0.5498613251580132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,256,1024,0.005076444397370021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,16384,0.6848097907172309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,12288,0.32105154461330837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,10240,0.4580071237352159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,12288,0.48413780000474715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,10240,0.2571217748853895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,8192,0.3662942250569661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,8192,0.19769599702623156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,10240,0.40593600273132324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,7168,0.3217128912607829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,8192,0.011965333587593503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,6144,0.2787777847713894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,8192,0.3457004494137234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,7168,0.2938008838229709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,6144,0.15202755398220485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,5120,0.23359465599060059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,7168,0.011016888750924004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,5120,0.14026310708787706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,4096,0.1901653342776828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,6144,0.24923200077480742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,3584,0.16539288891686335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,5120,0.212388449245029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,3584,0.09089066584904988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,4096,0.16666755411359999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,3072,0.14382755756378174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,3584,0.14058489269680446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,3072,0.08714133501052856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,2560,0.12093510892656113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,3072,0.12090578344133164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,2560,0.06874755356046888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,2048,0.09852888849046494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,2048,0.057688891887664795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,2560,0.09981599781248306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,1536,0.07627289162741767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,1536,0.0453342232439253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,2048,0.08227466874652438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,1024,0.05336533321274651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,1024,0.033554666572146945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,1536,0.06070755587683784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,768,0.0418853329287635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,1024,0.041257778803507485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,768,0.02804622385236952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,768,0.03272799981964959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,512,0.022574222750133936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,512,0.030928889910380047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,512,0.0270462218258116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,256,0.017106667160987854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,256,0.017875555488798354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,128,0.013144888811641268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,256,0.02225333286656274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,128,0.02033244404527876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,128,128,5120,0.008979555633332994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,64,0.008079999850855933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,128,0.010390222072601318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,65536,32,0.007359111474619971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,64,0.024302222662501868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,65536,32,0.02553244431813558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,16384,0.17845688925849068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,16384,0.09661600324842666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,16384,0.1579973300298055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,12288,0.08033777607811822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,12288,0.13616089026133218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,12288,0.11310756206512451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,65536,0.7318808767530652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,10240,0.06760088602701823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,10240,0.1142444478140937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,8192,0.09218933184941609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,10240,0.0932026637925042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,7168,0.04882133338186476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,8192,0.08118489053514269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,7168,0.08123199807273017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,65536,0.6296550962660048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,7168,0.06808088885413276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,6144,0.07534311215082805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,6144,0.059192001819610596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,5120,0.05987199809816149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,5120,0.03707199957635667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,5120,0.0508284436331855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,4096,0.0314631097846561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,4096,0.04898666673236423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,4096,0.04032177726427714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,3584,0.027648000253571406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,3584,0.04394666684998406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,3584,0.03637955586115519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,3072,0.03811733259095086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,3072,0.024567110670937434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,2560,0.021725333399242822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,3072,0.031143999762005273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,2560,0.032658666372299194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,2560,0.026269333230124578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,2048,0.027016000615225896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,2048,0.0221377776728736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,1536,0.015848888291252982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,1536,0.02160533269246419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,1536,0.018185777796639335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,1024,0.012835555606418185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,1024,0.015182221929232279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,1024,0.014236445228258768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,768,0.011379555695586734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,768,0.012016889121797351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,768,0.012816000315878125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,512,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,512,0.009019555317031013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,512,0.011519999967681037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,256,0.006451555424266391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,256,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,256,0.008365333080291748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,128,0.0058257778485616045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,128,0.00423555572827657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,64,0.003928889003064898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,128,0.007733333441946242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,16384,32,0.004129777765936322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,64,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,16384,32,0.009516444471147325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,65536,0.5399209128485786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,65536,0.3024657832251655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,7168,0.16463732719421387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,16384,0.07705689138836332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,65536,4096,0.10130133231480916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,16384,0.1491146617465549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,16384,0.11645510461595322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,12288,0.06045866674847073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,12288,0.10525421963797675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,65536,0.47433869043986004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,12288,0.08480444219377305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,10240,0.0906897783279419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,10240,0.0510391096274058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,8192,0.07169244686762492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,10240,0.07171644767125447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,8192,0.04221866528193156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,8192,0.0603075557284885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,7168,0.06413688924577501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,7168,0.052737779087490506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,6144,0.05616355604595608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,6144,0.033899555603663124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,6144,0.04498933421240913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,5120,0.047114666965272695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,5120,0.02921600143114726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,4096,0.03933511177698771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,5120,0.03943199912707011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,4096,0.024710221423043147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,4096,0.03205688794453939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,3584,0.03347377644644843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,3584,0.02499733368555705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,3584,0.02840977907180786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,3072,0.020096888144810993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,3072,0.029655999607510034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,2560,0.025482666161325242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,3072,0.024915556112925213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,2560,0.018055111169815063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,2048,0.021174222230911255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,2560,0.021506667137145996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,2048,0.01570222278436025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,2048,0.01812533371978336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,1536,0.013551111022631327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,1536,0.019155555301242404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,1536,0.014959999256663851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,1024,0.01107377807299296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,1024,0.0120328888297081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,1024,0.012527111503813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,768,0.009874666730562845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,768,0.010048000348938836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,768,0.011225777367750803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,512,0.00774133370982276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,512,0.008824889030721452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,512,0.009714666340086196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,256,0.006301333506902059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,256,0.004939555707905027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,256,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,128,0.005232000102599462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,128,0.0038337777886125776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,128,0.00700711127784517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,64,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,64,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,12288,32,0.0037644443412621817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,12288,32,0.008106666306654612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,8192,0.052556445201238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,65536,0.47236177656385636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,65536,0.35835909843444824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,16384,0.07688089211781819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,16384,0.14569421609242758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,65536,0.2560933430989583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,6144,0.041646222273508705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,12288,0.0964133342107137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,12288,0.05261866582764519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,16384,0.11014578077528213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,10240,0.07356711228688557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,10240,0.04486666785346138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,12288,0.07627466652128431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,8192,0.05921955241097344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,10240,0.0688248872756958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,8192,0.03756711218092177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,7168,0.03381422162055969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,7168,0.05265511075655619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,8192,0.05731022357940674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,65536,0.4514560169643826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,6144,0.04676444331804911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,6144,0.02978488802909851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,5120,0.025547555751270715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,6144,0.04299822118547228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,7168,0.05009511113166809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,5120,0.035671111610200666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,5120,0.04792622062895033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,4096,0.03240888979699876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,4096,0.02160800000031789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,4096,0.028896000650193956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,3584,0.029083556599087183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,3072,0.025872000389628943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,3584,0.019906666543748643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,3072,0.017768889665603638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,3584,0.026921777261628047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,2560,0.02236088944805993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,3072,0.023287110858493384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,2560,0.016186666157510545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,2560,0.020211555891566806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,2048,0.01847022275129954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,2048,0.013979555832015144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,1536,0.014488000008794995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,2048,0.016691555579503376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,1536,0.012264889147546558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,1024,0.010595555934641095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,1536,0.01426666643884447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,1024,0.012080888781282636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,768,0.008938666847017076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,768,0.009120889008045197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,768,0.01075644459989336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,512,0.00810933361450831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,512,0.008848889006508721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,512,0.007876444194051955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,256,0.004318222403526306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,256,0.007384888827800751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,128,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,128,0.0038053331275780997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,64,0.0034168888297345904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,128,0.005070222334729301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,16384,2048,0.018896889355447557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,32,0.007790221936172909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,10240,32,0.003573333223660787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,10240,64,0.007456888755162557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,16384,0.10022488567564224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,16384,0.07649066713121203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,65536,0.3820462226867676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,12288,0.08199644751018949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,12288,0.04244177871280246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,65536,0.1977004475063748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,12288,0.05769511063893636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,10240,0.0358897778722975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,65536,0.31414932674831814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,10240,0.06634577777650622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,8192,0.030113776524861652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,10240,0.05003822180959913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,8192,0.05570844146940443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,8192,0.041506667931874595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,7168,0.04257422354486254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,7168,0.027070222629441157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,6144,0.03788266579310099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,7168,0.03609066539340549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,6144,0.024110222856203716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,6144,0.03153422143724229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,5120,0.032103111346562706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,5120,0.020995555652512446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,5120,0.027448889282014634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,4096,0.026300445199012756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,4096,0.018245332770877414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,4096,0.022414222359657288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,3584,0.023473777704768713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,3584,0.01680888897842831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,3072,0.02086666723092397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,3584,0.02035999960369534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,3072,0.015218666858143277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,3072,0.018481777773963082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,2560,0.01795733306143019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,2560,0.013765333427323235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,2560,0.015839111473825242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,2048,0.014762666490342883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,2048,0.012281777958075205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,2048,0.013707555830478668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,1536,0.01161688897344801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,1536,0.010637332995732626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,1536,0.011805333197116852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,1024,0.008825778133339351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,1024,0.010093332992659675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,1024,0.009202666580677032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,768,0.007623111208279927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,768,0.008520888785521189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,768,0.00905777762333552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,512,0.0063528890411059065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,512,0.006050666587220297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,512,0.007148444652557373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,256,0.004045333299371931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,256,0.00654666663871871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,128,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,128,0.0048142220411035745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,128,0.006172444257471297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,64,0.0033155555526415506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,64,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,8192,32,0.0035866668654812705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,8192,32,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,65536,0.33905778990851504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,65536,0.1900764438841078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,16384,0.04866666595141093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,12288,7168,0.038282665941450335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,16384,0.08952977922227646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,12288,0.03865955604447259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,65536,0.30131377114189994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,16384,0.06844533152050443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,12288,0.06243111027611626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,10240,0.05265688896179199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,12288,0.05435199870003594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,10240,0.04790488878885905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,8192,0.042555554045571216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,8192,0.027599111199378967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,7168,0.03792266713248359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,7168,0.024978665841950312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,8192,0.041166222757763334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,6144,0.033915556139416166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,7168,0.03595644566747878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,6144,0.0221377776728736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,5120,0.028655999236636694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,5120,0.019575999842749704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,6144,0.03040177623430888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,4096,0.023688889212078516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,4096,0.01726222203837501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,5120,0.026532444689008925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,4096,0.021356445219781663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,3584,0.02128355536195967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,3584,0.015824000040690105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,3072,0.018851555056042142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,3584,0.018772444791264005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,3072,0.01761777698993683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,2560,0.013079110946920184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,2560,0.016200888488027785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,2560,0.015255999233987598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,2048,0.011721777419249216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,2048,0.013198222551080914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,2048,0.013638222383128272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,1536,0.010452444354693094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,1536,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,1536,0.011651555697123209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,1024,0.009158222211731805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,1024,0.008275555239783393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,1024,0.010056888891590966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,768,0.007128000259399414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,768,0.008126222425036961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,768,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,512,0.006115555349323485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,512,0.005695999910434087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,512,0.007228444847795699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,256,0.0038151111867692736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,256,0.0052284445199701525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,256,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,128,0.004861333303981357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,128,0.0034391110142072043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,128,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,64,0.0034417777011791864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,64,0.00666133314371109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,7168,32,0.007082666787836287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,7168,32,0.0034684443639384378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,65536,0.2813075648413764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,65536,0.17845243877834746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,16384,0.04914222160975138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,65536,0.2706106768714057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,16384,0.08056533336639404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,256,0.005446222093370225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,16384,0.06747466988033719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,10240,1024,0.010275555153687796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,12288,0.0542719993326399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,12288,0.03885866536034478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,12288,0.054656889703538686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,10240,0.03361777795685662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,10240,0.04595288965437147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,8192,0.03693155447642008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,8192,0.027657777070999146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,10240,0.04674933354059855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,16384,0.05311555663744608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,7168,0.024971556332376268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,8192,0.040552000204722084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,7168,0.03328355484538608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,7168,0.03552533189455668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,6144,0.029719110992219713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,6144,0.022256000174416438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,5120,0.019388443893856473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,6144,0.0295688874191708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,5120,0.025397333833906386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,5120,0.024891555309295654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,4096,0.021075555019908484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,4096,0.01661155621210734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,4096,0.020640000700950623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,3584,0.01868266695075565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,3584,0.015633778439627755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,3584,0.018252443936136033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,3072,0.016639999217457242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,3072,0.01404711107412974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,3072,0.016523554921150208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,2560,0.014179555906189812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,2560,0.012884444660610624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,2560,0.014279110564125909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,2048,0.012223111258612739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,2048,0.011358222199810876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,2048,0.012736000120639801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,1536,0.00981422182586458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,1536,0.011572444604502784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,1024,0.007911110917727152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,1024,0.008616888688670265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,1024,0.009333333207501305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,768,0.0068862222962909276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,768,0.007770666645632849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,768,0.006737777756320105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,512,0.0046693334976832075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,512,0.00720088928937912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,256,0.00488800017370118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,256,0.003730666720204883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,256,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,128,0.0033253334048721525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,128,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,64,0.0032213332338465583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,64,0.006676444576846228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,6144,32,0.0032968889507982465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,6144,32,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,65536,0.17647200160556367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,65536,0.2340471214718289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,8192,256,0.005167110926575131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,16384,0.0418542226155599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,16384,0.06807555754979451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,16384,0.0656800005171034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,65536,0.26288265652126735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,12288,0.05369777811898125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,10240,0.028355555401908025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,12288,0.04948888884650337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,10240,0.039459556341171265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,10240,0.04170133339034187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,8192,0.023597333166334365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,8192,0.03200799889034695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,8192,0.03492444422509935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,7168,0.02049688829316033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,7168,0.028632889191309612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,7168,0.030584888325797185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,6144,0.01813155578242408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,6144,0.02574933403068119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,6144,0.02680355476008521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,5120,0.021876444419225056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,5120,0.023139556248982746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,4096,0.013998222019937305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,4096,0.018034666776657104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,4096,0.019154666198624503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,3584,0.012968888713253869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,3584,0.016249777542220224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,3584,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,3072,0.012081777883900536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,3072,0.014412444498803882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,3072,0.015424888994958667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,2560,0.011147555377748279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,2560,0.012267555627557965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,10240,0.03340088989999559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,2560,0.013694222602579327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,2048,0.010433778166770935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,2048,0.010051555931568146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,2048,0.012328888807031842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,1536,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,1536,0.00868088917599784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,1536,0.010788444843557147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,1024,0.007107555866241455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,1024,0.006897777732875612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,1024,0.00850666645500395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,768,0.006021333237489064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,768,0.006090666684839461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,768,0.007888000044557784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,512,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,512,0.004265777766704559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,512,0.007087111473083496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,7168,3072,0.014337778091430664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,256,0.004960888789759742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,256,0.003597333199448056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,256,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,128,0.003269333392381668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,128,0.004726222405831019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,128,0.006208000083764394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,64,0.0030657777355776895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,64,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,5120,32,0.0031439999325407874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,5120,32,0.006800888727108638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,65536,0.18884978029463026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,65536,0.15452088250054255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,16384,0.03989244500796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,16384,0.05305244525273641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,65536,0.15731733375125462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,16384,0.04101777738995022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,12288,0.030096888542175293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,12288,0.04270311196645101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,12288,0.03116977877087063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,10240,0.03680977887577481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,10240,0.025996444953812495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,10240,0.02680977847841051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,8192,0.021990221407678392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,8192,0.03103733393881056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,8192,0.0229413327243593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,7168,0.01920622256067064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,7168,0.028343998723559912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,7168,0.020262221495310467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,6144,0.01685244507259793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,6144,0.025664000047577754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,6144,0.018040888839297824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,5120,0.018235555953449674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,5120,0.01629333363638984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,4096,0.012831110921170978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,4096,0.014966222974989148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,4096,0.01388711151149538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,3584,0.011891555454995898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,3584,0.013600000076823764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,3584,0.012883555557992725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,3072,0.010871999793582492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,3072,0.012058667010731168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,3072,0.012028444144460889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,2560,0.010085333552625444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,2560,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,2560,0.011028444601429833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,2048,0.009101333717505137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,2048,0.009115555220180089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,2048,0.010016000105275048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,1536,0.007785778078767989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,1536,0.00889866633547677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,1024,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,1024,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,1024,0.007048889166778988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,768,0.005531555662552516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,768,0.005072888814740711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,768,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,1536,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,512,0.0037564444873068067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,512,0.005378666851255629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,512,0.005172444300519096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,512,0.006173333360089197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,256,0.003415999934077263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,256,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,256,0.0057911111248864075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,128,0.003072000005178981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,128,0.004418666578001446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,128,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,64,0.0029848888516426086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,64,0.005838222387764189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,4096,32,0.0032017777363459268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,4096,32,0.006031999985376994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,6144,128,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,16384,0.03856711255179511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,65536,0.17423555586073133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,16384,0.0477182235982683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,16384,0.039106667041778564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,65536,0.15181510978274876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,65536,0.14644711547427708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,12288,0.03852977686458164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,12288,0.029832889636357624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,10240,0.03308799862861633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,10240,0.024315555890401203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,10240,0.025722665919197932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,8192,0.021480000681347314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,8192,0.028059555424584284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,7168,0.025563556287023757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,8192,0.021526222427686054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,7168,0.019380443625979953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,6144,0.01609866652223799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,6144,0.020394666327370536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,6144,0.017815111411942374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,5120,0.014561777313550314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,5120,0.016399999459584553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,5120,0.015467555986510383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,12288,0.030807998445298936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,4096,0.013462222284740873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,4096,0.013520888984203339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,5120,5120,0.016324443949593436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,3584,0.012244444754388599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,3584,0.012303111453851065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,3072,0.01035644445154402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,3072,0.011037333144081963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,3072,0.011592888997660743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,2560,0.009760888914267222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,2560,0.009744889206356471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,2560,0.010713777608341642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,2048,0.008729777402347988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,2048,0.008468444148699442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,2048,0.009726222190592024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,1536,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,1536,0.007345777418878343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,1536,0.00810933361450831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,1024,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,1024,0.005972444597217772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,1024,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,768,0.005511999958091312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,768,0.004571555389298333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,768,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,512,0.005136000199450387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,512,0.003755555384688907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,512,0.006175999840100606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,256,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,256,0.003351111171974076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,256,0.0058124446206622655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,128,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,128,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,64,0.0029973333908451926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,64,0.0058311112225055695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3584,32,0.0030177777840031516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3584,32,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,65536,0.14166577657063803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,65536,0.14560088846418592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,65536,0.12807822227478027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,16384,0.036743111080593534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,16384,0.04183377822240194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,16384,0.03696533375316196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,12288,0.02760800056987339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,12288,0.03420799970626831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,12288,0.02756799923049079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,10240,0.023736889163653057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,10240,0.029598222838507757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,10240,0.024089778463045757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,5120,0.01479555500878228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,8192,0.02027466727627648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,8192,0.02497333288192749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,8192,0.020398222737842135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,7168,0.017842666970358956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,7168,0.022845332821210224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,7168,0.01828444500764211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,6144,0.0173653331067827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,6144,0.016530666086408827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,4096,1536,0.008261333737108443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,5120,0.014656888114081489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,5120,0.01479555500878228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,12288,0.028311110205120508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,4096,0.012004444168673621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,4096,0.01220444424284829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,4096,0.012704888979593912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,3584,0.011167999770906238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,3584,0.01126933346192042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,3584,0.01180622229973475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,3072,0.010135999984211391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,3072,0.01000711073478063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,3072,0.01110400011142095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,2560,0.009053332938088311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,2560,0.010390222072601318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,7168,0.018538667096032035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,2048,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,2048,0.009212444225947062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,4096,0.012146666646003723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,1536,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,1536,0.007558222446176741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,1024,0.005695111221737332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,1024,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,1024,0.006780444333950679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,768,0.005423111220200856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,768,0.004283555679851108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,768,0.006456888798210356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,3584,0.011374221907721626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,512,0.0037786666717794207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,512,0.0061048890153567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,256,0.0046897778908411665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,256,0.0033768889390759995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,256,0.005762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,128,0.004480000171396467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,128,0.003142222141226133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,128,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,64,0.0029493332323100832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,64,0.0057706667317284485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,3072,32,0.002990222225586573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3584,128,0.004443555656406614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,3072,32,0.006008000009589725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,65536,0.13447022438049316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,65536,0.12391555309295654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,65536,0.12106755044725205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,16384,0.03515733281771342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,16384,0.03602933221393161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,16384,0.03338933322164748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,12288,0.027229333917299908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,12288,0.02926844358444214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,12288,0.02520711057715946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,10240,0.022286223040686712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,10240,0.025422222084469263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,10240,0.0222053329149882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,8192,0.020015999674797058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,8192,0.021799110703998145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,8192,0.01861422260602315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,7168,0.0176862229903539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,7168,0.019710222880045574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,7168,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,6144,0.016338666280110676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,6144,0.017246221502621967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,6144,0.015331554743978711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,5120,0.013780444032616086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,5120,0.015094222293959724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,5120,0.013843555417325763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,4096,0.01110400011142095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,4096,0.011770666473441653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,4096,0.011782222323947482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,3584,0.010981332924630908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,3584,0.011354666617181567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,3584,0.011218667030334473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,3072,0.009815110928482479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,3072,0.009391999906963771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,3072,0.010436444646782346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,2560,0.009001777403884465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,2560,0.008433777425024245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,2560,0.009752000371615091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,2048,0.0075502221783002215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,2048,0.0073520003093613525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,2048,0.00851288851764467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,1536,0.006392000036107169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,1536,0.006528000036875407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,1536,0.007347555624114142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,1024,0.0047937776479456155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,1024,0.005751111027267244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,1024,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,768,0.0042408886882993905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,768,0.0063182223174307085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,512,0.00498844434817632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,512,0.003673777812057071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,6144,0.015916445189052157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,512,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,256,0.0032924444725116095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,256,0.004625777817434735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,256,0.005669333454635408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,128,0.003019555575317807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,128,0.004426666845877965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,128,0.005479111025730769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,64,0.0029004443850782183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,64,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2560,32,0.0029191111938820947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2560,32,0.005756444401211209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,5120,0.014175999495718213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,65536,0.10189688867992824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,65536,0.11754933993021648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,16384,0.031146665414174397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,65536,0.08761066860622829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,16384,0.030237333642111883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,16384,0.0262044436401791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,12288,0.024751999311976965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,12288,0.02122933334774441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,10240,0.0206631107462777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,10240,0.02176355487770504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,10240,0.019705777366956074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,8192,0.018043556147151522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,8192,0.018162666095627677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,8192,0.01625955601533254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,7168,0.016064888901180692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,7168,0.016384000579516094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,7168,0.015217777755525378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,6144,0.014656000667148165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,6144,0.014302222265137566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,6144,0.014005333185195923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,5120,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,5120,0.012660444610648684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,4096,0.01094044413831499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,4096,0.010764444039927589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,4096,0.011013333168294696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,3584,0.010324444207880232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,3584,0.009878222313192155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,3584,0.010617777705192566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,3072,0.009213333328564962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,3072,0.009202666580677032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,3072,0.009761778016885122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,2560,0.008040889269775814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,2560,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,2560,0.008414222134484185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,2048,0.006723555425802867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,2048,0.007460444337791867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,2048,0.008284444610277811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,2048,0.007839110990365347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,1536,0.006115555349323485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,1536,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,1536,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,1536,0.007064888874689738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,1024,0.005438222239414851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,1024,0.00452533322903845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,1024,0.0064284445510970215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,768,0.003963555726740095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,768,0.006174222048785951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,512,0.004945777770545748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,512,0.0035946667194366455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,512,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,2560,0.009282666775915358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,256,0.0031742221779293488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,3072,512,0.005115555392371284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,256,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,128,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,128,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,64,0.002801777794957161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,64,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,2048,32,0.0029351111087534162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,2048,32,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,65536,0.09093333615197076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,65536,0.08014666371875338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,65536,0.08587200111813015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,16384,0.028155555327733357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,16384,0.024937777055634394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,16384,0.02535111043188307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,12288,0.020210666788948905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,12288,0.02089866664674547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,10240,0.018485332528750103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,10240,0.017882666654056974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,10240,0.018049778209792245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,8192,0.017238222890430026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,8192,0.015417777829700045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,8192,0.015601777368121676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,7168,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,7168,0.013883555928866068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,7168,0.014721777704026965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,6144,0.013734222286277346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,6144,0.012596444123321109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,6144,0.013527111046844058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,5120,0.012227555943859948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,5120,0.01109955542617374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,5120,0.012359110845459832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,4096,0.0099973330895106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,4096,0.00960000024901496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,4096,0.011018666956159802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,3584,0.009326222042242685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,3584,0.008928889201747047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,3584,0.010323555933104621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,3072,0.008408000071843466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,3072,0.008179555336634317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,3072,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,2560,0.007391110890441471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,2560,0.0074720001882976955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,2560,0.008283555507659912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,2048,0.006516444600290722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,2048,0.006548444430033366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,2048,0.0075528886583116316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,1536,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,1536,0.005212444398138259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,1536,0.007032000356250339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,1024,0.005370666583379109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,1024,0.004453333301676644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,1024,0.00646311127477222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,768,0.005106666849719154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,768,0.004061333421203825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,768,0.006213333457708359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,512,0.0047626665068997275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,512,0.0036071112586392295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,512,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,256,0.004446222136418025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,256,0.003168888803985384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,256,0.005639111002286275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,128,0.004283555679851108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,128,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,128,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,64,0.002776888923512565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,64,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1536,32,0.002850666642189026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1536,32,0.005655111124118169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,65536,0.09189066621992324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,65536,0.06234311395221286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,65536,0.07955555783377753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,16384,0.026363554928037856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,16384,0.018668444620238412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,16384,0.024352888266245525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,12288,0.021133333444595337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,12288,0.015267555912335714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2560,768,0.005371555685997009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,12288,0.01940355532699161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,10240,0.01791288952032725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,10240,0.014456889695591398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,10240,0.017464889420403373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,8192,0.016185777054892648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,8192,0.013427555561065674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,8192,0.015135111080275642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,7168,0.014295111099878946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,7168,0.012802666260136498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,7168,0.014018666413095264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,12288,0.0248355550898446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,6144,0.01197866681549284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,6144,0.013015111287434896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,5120,0.012655999925401477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,5120,0.01090133355723487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,5120,0.01200977795653873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,4096,0.009106666677527958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,4096,0.009758222434255812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,4096,0.00962755580743154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,3584,0.008569777839713627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,3584,0.00853066643079122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,3584,0.009003555609120263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,3072,0.007816000117195977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,3072,0.007843555675612556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,3072,0.008744888835483128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,2560,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,2560,0.006693333387374878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,2560,0.007881777981917063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,2048,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,2048,0.005479111025730769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,2048,0.007380444142553542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,1536,0.00582755563987626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,1536,0.004784000002675586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,1536,0.006830222076839871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,768,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,1024,0.004138666722509596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,1024,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,1024,0.0063902222447925145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,768,0.005087999833954705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,768,0.003768889026509391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,256,0.00452533322903845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,768,0.006093333164850871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,512,0.0034266666819651923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,512,0.005873777800136142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,256,0.003088888815707631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,256,0.004446222136418025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,256,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,128,0.004238222208287981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,128,0.0028533333291610083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,128,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,64,0.0027111110587914786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,64,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,1024,32,0.0027253333893087176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,1024,32,0.005501333210203383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,2048,128,0.004338666796684265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,65536,0.0521360006597307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,65536,0.07664977841907077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,65536,0.07218844360775419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,16384,0.01752622259987725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,16384,0.022589332527584497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,12288,0.014699555105633207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,12288,0.019288889235920377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,12288,0.018639999959203932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,10240,0.013245333400037555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,10240,0.016338666280110676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,10240,0.01681422193845113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,8192,0.015377778146002026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,8192,0.014533332652515836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,8192,0.011728000309732227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,7168,0.01368533323208491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,7168,0.011049778097205691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,7168,0.013616888887352414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,6144,0.010323555933104621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,6144,0.011556444068749746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,6144,0.012531555361217923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,5120,0.009657777845859528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,5120,0.010173333187898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,5120,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,4096,0.008731555607583787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,4096,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,4096,0.009304000271691216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,3584,0.008271999657154083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,3584,0.00812000036239624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,3584,0.00868800034125646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,3072,0.007361777954631382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,3072,0.0075191110372543335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,3072,0.008284444610277811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,2560,0.007807999849319458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,2560,0.0064693333374129395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,2048,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,2048,0.0054560001525614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,2048,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,1536,0.005744888964626525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,1536,0.004775111046102312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,1536,0.006787555499209299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,1024,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,1024,0.004085333396991094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,1024,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,768,0.003755555384688907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,768,0.005000888887378904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,768,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,512,0.0047306666771570844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,512,0.0034382223255104492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,512,0.0058053334554036455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,256,0.004407111141416761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,256,0.003068444422549672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,256,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,128,0.004201777693298128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,128,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,128,0.005381333331267039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,64,0.002704000100493431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,64,0.005468444277842839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,768,32,0.0027297778675953546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,768,32,0.005495111147562663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,65536,0.07044355736838447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,65536,0.04011822077963088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1536,12288,0.022516444325447083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,65536,0.06935999790827434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,16384,0.013768889009952545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,16384,0.021960000197092693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,12288,0.011933333343929715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,12288,0.018152889278199937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,10240,0.014845333165592618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,10240,0.010978666444619497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,10240,0.01633777717749278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,8192,0.012219555675983429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,8192,0.010021333065297868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,8192,0.013998222019937305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,7168,0.011134222149848938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,7168,0.009445333646403419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,7168,0.012897777888509961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,6144,0.009983110758993361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,6144,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,6144,0.011400000088744693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,5120,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,5120,0.008266666697131263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,5120,0.010288889209429422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,4096,0.008094222181373173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,4096,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,4096,0.009031111167536842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,3584,0.007541333635648091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,3584,0.007083555890454187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,3584,0.008591111335489485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,3072,0.007101332975758447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,3072,0.006704000135262807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,3072,0.0081395556529363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,2560,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,2560,0.006177777631415262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,2560,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,2048,0.006186666587988536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,2048,0.00544888898730278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,2048,0.007188444336255391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,1536,0.005688000056478713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,1536,0.004781333522664176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,1536,0.006743999818960826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,1024,0.005208000126812193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,1024,0.004077333129114575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,1024,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,768,0.004920000003443824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,768,0.0037271111375755737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,768,0.006058666855096817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,512,0.00462400002612008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,512,0.0034008889148632684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,512,0.0057857777509424425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,256,0.004328888737493091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,256,0.003035555490189128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,256,0.005487999982304043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,128,0.004211555752489301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,128,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,128,0.005362666729423735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,64,0.0026782221264309357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,64,0.005371555685997009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,512,32,0.0027635554886526535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,512,32,0.005381333331267039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,65536,0.070104890399509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,65536,0.02622844444380866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,65536,0.06802844338946872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,16384,0.020797333783573575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,16384,0.010413332945770687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,16384,0.021918222308158875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,12288,0.016029333074887592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,12288,0.009393778112199571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,12288,0.017337777548366122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,10240,0.013783111340469785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,10240,0.008845333423879411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,10240,0.01512977812025282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,8192,0.011894221935007306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,8192,0.007511999871995714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,8192,0.012676444318559436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,7168,0.010858666565683154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,7168,0.007350222104125553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,6144,0.0127368892232577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,7168,0.011659555964999728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,6144,0.006490666833188798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,6144,0.009866666462686326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,6144,0.010786666638321347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,5120,0.008984000318580205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,5120,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,5120,0.010071111222108206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,4096,0.007853333320882585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,4096,0.006526222245560751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,4096,0.008906666603353288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,3584,0.0068808889223469635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,3584,0.0075235557225015424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,3584,0.008567999634477828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,3072,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,3072,0.0064897777305708984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,3072,0.008185777399275038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,2560,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,2560,0.005993777679072485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,2560,0.007705777883529663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,2048,0.006069333189063602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,2048,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,2048,0.007119110888904995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,5120,0.011260444091426002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,1536,0.0047022220161226065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,1536,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,1024,0.005168888717889786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,1024,0.004001777619123459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,1024,0.00620266670982043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,1024,512,0.004851555658711327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,768,0.0036319999231232535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,768,0.006037333359320958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,512,0.003350222276316749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,512,0.005745777653323279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,256,0.004326222257481681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,256,0.0029813332690132987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,256,0.005469333380460739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,16384,0.02458488941192627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,128,0.004152889053026835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,128,0.005335111171007156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,128,0.002790222151411904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,64,0.0026213334252436957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,64,0.005377777748637729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,256,32,0.002630222174856398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,256,32,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,65536,0.017998221847746108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,65536,0.06993689139684041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,65536,0.06839999887678358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,16384,0.020492444435755413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,16384,0.008301333420806462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,16384,0.021431111627154883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,12288,0.007096000015735626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,12288,0.015628443823920358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,12288,0.016536000702116225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,10240,0.00721599989467197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,10240,0.014396443963050842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,8192,0.011764444410800934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,8192,0.006802666518423293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,8192,0.012224888636006249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,7168,0.010800888968838586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,7168,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,7168,0.011352889239788055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,768,2560,0.006894222150246303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,6144,0.009810667071077559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,6144,0.00638133328821924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,6144,0.010590222146775989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,5120,0.006793777975771163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,5120,0.008842666943868002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,5120,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,4096,0.007880888879299164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,4096,0.0063999998900625445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,4096,0.008647999829716152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,3584,0.007147555549939473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,3584,0.0074444446298811175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,3584,0.008185777399275038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,3072,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,3072,0.006507555643717448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,3072,0.007868444754017724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,2560,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,2560,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,2560,0.0074026667409473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,2048,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,2048,0.005261333452330695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,2048,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,1536,0.00544888898730278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,1536,0.0046471113132105935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,1536,0.006495999793211619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,1024,0.005029333134492238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,1024,0.003958222352796131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,1024,0.006081777728266186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,768,0.004783111313978831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,768,0.003645333151022593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,768,0.005869333528810077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,512,0.0032915555768542816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,512,0.004552000098758274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,512,0.005653333332803514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,256,0.004306666553020477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,256,0.0029644444584846497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,256,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,128,0.004085333396991094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,128,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,64,0.0026373333401150177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,128,0.0052879999081293745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,128,32,0.0026462222966882917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,64,0.0052986666560173035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,128,32,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,65536,0.014669333895047506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,16384,0.006826666494210561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,65536,0.06853689087761773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,16384,0.020729778541458976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,12288,0.0064702220261096954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,12288,0.01572177807490031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,10240,0.006694222076071634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,10240,0.013940445250935025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,8192,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,7168,0.006480000085300869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,8192,0.012093333734406365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,7168,0.01125600023402108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,6144,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,6144,0.010478222535716163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,5120,0.006708444406588872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,5120,0.009658666948477427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,4096,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,4096,0.008649778034951951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,3584,0.006762666834725275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,16384,0.020920000142521326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,3584,0.008223999705579545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,3072,0.007877333296669854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,3072,0.006412444429265128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,2560,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,2560,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,2048,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,2048,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,1536,0.004582222137186262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,1024,0.003899555653333664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,1536,0.006505777852402792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,1024,0.006063999815119638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,768,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,768,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,512,12288,0.016872889465755887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,512,0.0032275555034478507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,512,0.005659555395444234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,256,0.005434666656785541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,128,0.0027537778433826235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,256,0.00291377781993813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,128,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,64,0.002601777720782492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,64,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,64,32,0.0025537777692079544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,64,32,0.0053048887186580235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,65536,0.013417777915795645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,16384,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,12288,0.006380444599522485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,65536,0.06863911284340753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,16384,0.020869332883093093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,10240,0.006559111177921295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,12288,0.015621332658661736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,10240,0.013893332746293811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,8192,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,8192,0.012118221984969245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,7168,0.0064071110553211635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,7168,0.01128355579243766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,6144,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,6144,0.010468444062603844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,5120,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,5120,0.009697777529557547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,4096,0.006229333165619109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,4096,0.008644444247086843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,3584,0.006507555643717448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,3584,0.008231999973456064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,3072,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,2560,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,3072,0.007851555943489075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,2560,0.00738844441043006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,2048,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,2048,0.006903999795516332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,1536,0.004585777719815572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,1536,0.006503110958470239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,1024,0.003947555604908201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,1024,0.006062222023804982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,768,0.003588444242874781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,768,0.0058355554938316345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,512,0.0032497776879204642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,512,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,256,0.002941333378354708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,256,0.00544799988468488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,128,0.002720000015364753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,128,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,64,0.0025555555605226094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,64,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,96,32,32,0.0025742221623659134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,96,32,32,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,16384,0.7413893275790744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,16384,0.38138487603929305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,16384,0.7040293481614855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,12288,0.5712053510877821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,12288,0.32050222820705837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,10240,0.4741519822014703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,10240,0.24674577183193633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,12288,0.5070248709784614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,10240,0.4443991184234619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,8192,0.3825795650482178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,8192,0.21634399890899658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,7168,0.3330880006154378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,8192,0.3625644312964545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,7168,0.16340088844299316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,6144,0.2864888774024116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,7168,0.3116035461425781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,5120,0.23933598730299208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,6144,0.26990222930908203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,5120,0.1396933396657308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,4096,0.19465333885616729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,4096,0.10223999950620864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,5120,0.23024710019429526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,1536,0.005514666852023866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,768,0.004826666580306159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,3584,0.1720648871527778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,4096,0.18289866712358263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,3584,0.09386222230063544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,3584,0.15520000457763672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,3072,0.14798311392466226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,3072,0.08500177992714776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,2560,0.12497777409023708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,3072,0.13232978185017905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,256,512,0.004571555389298333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,2560,0.06732266479068331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,2048,0.09946399927139282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,2048,0.056976000467936196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,2560,0.11398400200737847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,1536,0.0785706639289856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,2048,0.08630755874845718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,1536,0.04452622267935011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,1024,0.0538257790936364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,1024,0.03321955601374308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,1536,0.06413688924577501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,1024,0.04378755556212532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,768,0.04356266723738777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,768,0.03491466575198703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,512,0.02203911046187083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,512,0.03108888864517212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,512,0.026625777284304302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,256,0.018530666828155518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,256,0.016985777351591323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,256,0.022270222504933674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,128,0.01107022249036365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,64,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,128,0.02034133341577318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,64,0.025331555141343012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,65536,32,0.007130666739410824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,65536,32,0.02586577832698822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,65536,0.7226399845547147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,16384,0.17909333440992567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,96,128,10240,0.01370933320787218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,16384,0.11124532752566868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,65536,0.38701687918768984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,12288,0.07669599850972493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,12288,0.13604711161719427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,10240,0.11374933189815944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,16384,0.17018755276997885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,65536,0.727567089928521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,10240,0.06478577852249146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,12288,0.13086399767133924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,8192,0.05233066611819797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,8192,0.09780089060465495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,10240,0.10855378044976129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,7168,0.08127466837565105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,7168,0.0487022234333886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,8192,0.09143200185563828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,6144,0.04116355710559421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,6144,0.07129422161314222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,7168,0.08018489016426934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,5120,0.05948978000217014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,6144,0.07022666931152344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,4096,0.04862577716509501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,5120,0.057215107811821826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,4096,0.03349155518743727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,3584,0.027160889572567407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,3584,0.04338488976160685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,4096,0.04598755637804667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,3584,0.04147288865513272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,3072,0.03803822067048814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,3072,0.0351582235760159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,2560,0.03228177626927694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,3072,0.027172444595230952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,2560,0.02167822255028619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,2048,0.026643556025293138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,2560,0.029771556456883747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,2048,0.024195555183622573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,1536,0.021171554923057556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,1536,0.019344889455371432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,1024,0.012606222596433429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,1024,0.015113777584499784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,1024,0.0151555554734336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,768,0.011228444675604502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,768,0.011797332929240333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,768,0.013353777428468069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,512,0.009745777481132084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,512,0.008900444540712569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,512,0.01181777815024058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,256,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,256,0.00868266655339135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,256,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,128,0.0041697778635554844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,128,0.005689777847793367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,128,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,64,0.0038195554580953387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,16384,32,0.004026666697528628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,64,0.009352000223265754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,16384,32,0.009679111341635386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,6144,0.1430035564634535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,65536,0.5382613076104058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,16384,0.08854577938715617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,65536,0.3007413281334771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,16384,0.14785688453250462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,12288,0.059824003113640674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,16384,0.12681244479285345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,12288,0.11045333411958481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,65536,0.5097484588623047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,12288,0.09689778089523315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,10240,0.09357777569029067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,10240,0.0820524427625868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,8192,0.042189333173963756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,768,0.029438220792346533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,8192,0.07570755481719971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,65536,128,0.013372444444232516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,7168,0.0667111078898112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,8192,0.06863288746939765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,7168,0.0377768874168396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,6144,0.05795733133951823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,7168,0.059664891825781934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,6144,0.033731556600994535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,6144,0.05132711264822218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,5120,0.04833955566088358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,5120,0.02910311023394267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,4096,0.040039112170537315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,4096,0.02411288850837284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,5120,0.04343733191490173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,3584,0.035623110002941556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,3584,0.022321777211295232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,4096,0.03484800126817491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,3072,0.030830221043692693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,3584,0.032074666685528226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,3072,0.027635556128289964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,2560,0.017895999881956313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,2560,0.02649066680007511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,2560,0.023440000083711412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,2048,0.015448000695970325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,2048,0.02223377757602268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,2048,0.019308444526460435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,1536,0.013194666968451606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,1536,0.017449777987268236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,1536,0.015967110792795818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,1024,0.012704888979593912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,1024,0.012798222402731577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,768,0.010453333457310995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,768,0.011621333658695221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,768,0.009950222240553962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,512,0.008742222355471717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,512,0.009775111244784461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,512,0.007986666427718269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,256,0.005023111071851519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,256,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,256,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,128,0.005163555757866966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,128,0.003740444365474913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,128,0.007093333535724216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,64,0.00342755557762252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,64,0.008242666721343994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,12288,32,0.0036728887094391715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,12288,32,0.008268444074524773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,65536,0.468190246158176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,5120,0.035629332065582275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,65536,0.2503111097547743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,2048,0.01867466668287913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,16384,0.1300533347659641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,16384,1536,0.015964445140626695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,12288,0.058879110548231334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,16384,0.10993066761228774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,12288,0.08807822068532307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,12288,0.08794933557510376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,10240,0.0743031104405721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,65536,0.4890720049540202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,10240,0.07276266813278198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,8192,0.03894133369127909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,8192,0.06868000162972344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,8192,0.06141777833302816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,7168,0.052779555320739746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,7168,0.050887111160490245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,6144,0.029548443026012842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,6144,0.046816888782713145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,6144,0.04587999979654948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,5120,0.025599110457632277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,5120,0.03936800029542711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,5120,0.040080888403786555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,4096,0.021228444245126512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,4096,0.03258400162061056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,4096,0.033414221472210355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,3584,0.019719999697473314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,3584,0.0288426677385966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,3584,0.02940800123744541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,3072,0.01792622274822659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,3072,0.025614221890767414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,3072,0.024917332662476435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,2560,0.01609599921438429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,2560,0.021927111678653296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,2560,0.02232622272438473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,2048,0.013923555612564087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,2048,0.018214222457673814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,2048,0.018816888332366943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,1536,0.012292444705963135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,1536,0.014300444059901767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,1536,0.014954666296641031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,1024,0.010144000252087912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,1024,0.010543111297819348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,1024,0.012231999801264869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,768,0.009159111314349705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,768,0.008880889250172509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,768,0.011041777829329172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,512,0.007713778151406183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,512,0.00739911115831799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,512,0.00906488878859414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,256,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,256,0.004400889078776042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,256,0.0075528886583116316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,128,0.00499555551343494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,128,0.0037368887828456033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,128,0.007081777685218387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,64,0.0034613334056403902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,64,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,10240,32,0.0034844444857703317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,10240,32,0.007844444778230455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,10240,0.05078666739993625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,65536,0.23262222607930502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,65536,0.37126933203803164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,16384,0.09984622399012248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,16384,0.08508977625105117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,65536,0.3456302218967014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,12288,0.042837331692377724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,3072,0.019935111204783123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,12288,0.07064444488949247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,12288,1024,0.010995555255148145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,10240,0.058416889773474805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,10240,0.03630133469899496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,12288,0.06502488586637709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,8192,0.02986933456526862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,10240,0.05587733454174466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,8192,0.0476382209195031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,8192,0.046091556549072266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,7168,0.04219111137919956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,7168,0.026887999640570745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,7168,0.04048177931043837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,6144,0.037447998921076454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,6144,0.023921777804692585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,6144,0.03544355432192484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,5120,0.031567109955681696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,5120,0.030637333790461223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,4096,0.026124444272783067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,4096,0.018265777164035372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,3584,0.023392000132136877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,4096,0.024322667055659827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,3584,0.016910221841600206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,3072,0.0206675562593672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,3072,0.015132443772421943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,3584,0.022634666826989915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,3072,0.019849777221679688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,2560,0.017846221725145977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,2560,0.013757333159446716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,2560,0.017331555485725403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,2048,0.014535999960369535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,2048,0.01219733307758967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,2048,0.01442577772670322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,1536,0.01163999984661738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,1536,0.010619555910428366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,1536,0.012368000215954252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,1024,0.008760000268618265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,1024,0.009149333669079674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,1024,0.010288000106811523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,768,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,768,0.007469333708286285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,768,0.008914666871229807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,512,0.006054222169849608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,512,0.007463110817803278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,512,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,256,0.003983111017280155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,256,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,256,0.00518488883972168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,128,0.0034302222645945023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,128,0.004757333132955763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,128,0.0063235556913746735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,64,0.0032995556377702286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,64,0.00685511115524504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,8192,32,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,8192,32,0.007063999772071838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,65536,0.34310134251912433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,16384,0.06616177823808458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,16384,0.05445422066582573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,65536,0.23855998780992296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,16384,0.08799911207622951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,65536,0.33185333675808376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,12288,0.06201689110861885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,12288,0.04084533452987671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,16384,0.0831164452764723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,10240,0.05215200119548374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,12288,0.06467644373575847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,10240,0.03492533498340183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,10240,0.053166220585505165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,8192,0.04236622320281135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,8192,0.030220445659425523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,10240,0.04457777738571167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,8192,0.04003999961747064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,7168,0.037610666619406805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,7168,0.025763556361198425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,6144,0.03348799877696567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,7168,0.03833599885304769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,5120,0.02830933199988471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,6144,0.032700445916917585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,5120,0.01964977714750502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,5120,0.027258666025267705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,4096,0.017794667018784415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,4096,0.023600000474188063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,4096,0.022453332940737408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,3584,0.02105866703722212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,3584,0.01572800013754103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,3584,0.020603555772039626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,3072,0.018596443865034316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,3072,0.014341332846217685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,3072,0.019202666150199044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,2560,0.016010666886965435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,2560,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,2048,0.013206221991115145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,2560,0.016965332958433364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,2048,0.011720889144473605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,10240,7168,0.03345333205329047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,1536,0.010670222342014313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,2048,0.014311111635631986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,1536,0.010397333237859938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,1536,0.012661333713266583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,1024,0.008223999705579545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,1024,0.009082666701740688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,1024,0.010148444109492833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,768,0.007038222418891058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,768,0.008982222113344405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,768,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,512,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,512,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,512,0.007238222493065729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,256,0.003670222229427762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,256,0.006684444430801604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,128,0.0033200000309281876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,128,0.004836444639497333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,128,0.006326222171386083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,64,0.006702222343948152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,64,0.0033564445459180405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,7168,32,0.0034595556143257353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,7168,32,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,65536,0.2809048758612739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,16384,0.05513066715664334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,65536,0.19705243905385336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,16384,0.07951111263699003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,65536,0.29841778013441295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,12288,0.05401688814163208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,16384,0.07562933365503947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,12288,0.05831288629108005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,10240,0.04554311103290982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,10240,0.05069955521159702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,8192,0.02773155603143904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,8192,0.03695644603835212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,8192,0.04169333312246534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,7168,0.02495999965402815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,7168,0.03288799855444167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,7168,0.03420889046457078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,6144,0.02201333310869005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,6144,0.02941688895225525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,6144,0.031621333625581526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,5120,0.019317333896954853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,5120,0.024996444582939148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,4096,0.016473778420024447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,5120,0.027543110979927912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,4096,0.02103022237618764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,4096,0.02199466692076789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,3584,0.01868355605337355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,3584,0.015400888191329109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,3584,0.01925155520439148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,3072,0.013970666461520724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,3072,0.016619554824299283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,3072,0.017934223016103108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,2560,0.014222222897741528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,2560,0.012662221988042196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,2560,0.01661777827474806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,2048,0.011200889117187925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,2048,0.012196444802814059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,2048,0.01349422252840466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,16384,0.05397955576578776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,1536,0.009886222581068674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,1536,0.009964444571071202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,1024,0.007855110698276097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,1536,0.01146311147345437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,1024,0.008430221842394935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,1024,0.009529777699046666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,768,0.006701333241330252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,768,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,768,0.007917332980367873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,512,0.005468444277842839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,512,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,512,0.0046302220887608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,256,0.004862222406599256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,256,0.003642666671011183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,256,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,128,0.003293333368168937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,128,0.004552888787455029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,128,0.006219555520349079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,64,0.003084444544381566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,64,0.006786666810512543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,6144,32,0.003129777808984121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,6144,32,0.006801777829726537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,65536,0.23361778259277344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,65536,0.1950835519366794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,65536,0.27749066882663304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,16384,0.048772444327672325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,16384,0.06710400183995564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,12288,0.03776622149679396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,16384,0.06714222166273329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,12288,0.05284622311592102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,12288,0.05291733476850721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,8192,5120,0.020900444851981267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,10240,0.03141511148876614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,10240,0.03875289029545254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,8192,0.03156266609827677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,10240,0.04638044370545281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,8192,0.02588266630967458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,7168,0.028296000427669946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,8192,0.03907288776503669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,7168,0.022797332869635686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,7168,0.03316266669167413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,6144,0.02514222264289856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,6144,0.020243555307388306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,6144,0.028732445504930284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,5120,0.021513778302404616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,5120,0.01773599949147966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,5120,0.024873778223991394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,4096,0.014876445134480795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,4096,0.01774311065673828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,3584,0.015966221690177917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,4096,0.02071822186311086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,3584,0.01344622257683012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,3072,0.014105778601434497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,3584,0.019792889555295307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,3072,0.012266666524940066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,3072,0.01676266723208957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,2560,0.011994666523403592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,2560,0.011177777416176267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,2048,0.010285333626800114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,2560,0.014894222219785055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,2048,0.009970666633711921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,2048,0.012686221963829465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,1536,0.008619555168681675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,1536,0.008991111483838823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,1536,0.01129155523247189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,1024,0.006829333388143116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,1024,0.007263999846246507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,1024,0.008609777523411645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,768,0.006046222315894232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,768,0.006043555421961679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,768,0.007888889147175683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,512,0.004146666576464971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,512,0.005527110977305307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,512,0.0070168889231152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,256,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,256,0.004931555440028508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,256,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,128,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,5120,128,0.004598222259018156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,128,0.006194666855865055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,64,0.003015999992688497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,5120,32,0.0030871110243929755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,64,0.006597333484225803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,5120,32,0.006700444552633498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,65536,0.1881999969482422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,16384,0.04351822204060025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,65536,0.16605510976579455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,65536,0.16903200414445665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,16384,0.05219377742873298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,16384,0.04451644420623779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,12288,0.041857779026031494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,12288,0.03341689043574863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,12288,0.03542222248183356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,10240,0.036088890499538846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,6144,0.022527111901177302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,10240,0.028288000159793433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,10240,0.03016888764169481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,8192,0.03035466538535224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,8192,0.023899555206298828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,8192,0.025096888343493145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,7168,0.027798222170935735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,7168,0.022389334109094407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,6144,0.018443556295500863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,6144,0.024920889072948035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,6144,0.020103999310069613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,5120,0.018230222993426852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,5120,0.016141333513789706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,4096,0.014942222171359591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,5120,0.01779022150569492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,4096,0.013639111485746173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,4096,0.01475822263293796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,3584,0.012629333469602795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,3584,0.013667555318938361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,3584,0.013440888788965015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,3072,0.011552888486120435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,3072,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,3072,0.012655999925401477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,2560,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,2560,0.010510221951537661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,2560,0.011325333681371478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,2048,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,2048,0.010085333552625444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,1536,0.008103999826643202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,1536,0.007722666694058313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,1536,0.00903200027015474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,1024,0.006106666806671355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,1024,0.007179555793603261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,768,0.005753777921199799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,768,0.005068444543414646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,768,0.00665511108107037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,512,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,512,0.0036862223512596558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,512,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,256,0.004729777574539185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,256,0.0033075554917256036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,256,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,7168,256,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,128,0.004436444491147995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,128,0.003087999920050303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,128,0.005639111002286275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,64,0.0029760001020299066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,64,0.006017777654859755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,4096,32,0.003208000005947219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,4096,32,0.006137777947717243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,65536,0.17294489012824166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,65536,0.15624088711208767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,65536,0.1529902219772339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,12288,0.03919644488228692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,16384,0.049136890305413135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,16384,0.04230133361286587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,12288,0.032595554987589516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,6144,10240,0.03357244531313578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,12288,0.03795377744568719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,12288,0.032925334241655134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,10240,0.026649778087933857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,10240,0.032778667079077825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,10240,0.028312000963422988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,8192,0.02757955590883891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,7168,0.020346666375796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,8192,0.02353511088424259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,7168,0.021220443977249995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,7168,0.02090222140153249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,6144,0.01791555517249637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,6144,0.019275556008021038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,6144,0.019150222341219585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,5120,0.016336000627941556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,5120,0.017044444878896076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,4096,0.013232889274756113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,4096,0.013385777672131857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,4096,0.014148443937301636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,3584,0.012317332956526013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,3584,0.012142221960756512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,3584,0.013183111117945777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,3072,0.011176000038782755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,3072,0.01088711122671763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,3072,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,2560,0.009713778065310584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,2560,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,2560,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,2048,0.008470222353935242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,2048,0.008723555339707268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,2048,0.009783999787436591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,1536,0.007453333172533248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,1536,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,1536,0.008602666358153025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,1024,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,1024,0.0059057776298787855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,1024,0.007072889142566257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,768,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,768,0.004354666504595014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,768,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,512,0.005186666631036335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,512,0.0036755556033717263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,512,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,256,0.0032995556377702286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,256,0.004736888739797804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,256,0.00582311095462905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,128,0.004390222330888112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,128,0.0030337776988744736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,128,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,64,0.002899555489420891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3584,32,0.0029768887907266617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,64,0.005850666513045629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3584,32,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,65536,0.14511199792226157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,65536,0.150583995713128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,16384,0.03971199856864081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,65536,0.1404213375515408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,16384,0.04433777928352356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,12288,0.030206223328908283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,16384,0.03929511043760512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,12288,0.033273776372273765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,12288,0.03035822179582384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,10240,0.025615110993385315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,10240,0.028654221031400893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,10240,0.026375111606385972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,8192,0.021543999512990315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,8192,0.024459555745124817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,8192,0.022039999564488728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,7168,0.019188443819681805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,7168,0.02204088866710663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,7168,0.019654222660594516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,6144,0.017163554827372234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,6144,0.01982933282852173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,6144,0.01791733337773217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,5120,0.015073777900801765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,5120,0.01731911136044396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,5120,0.015736889508035447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,4096,0.012627555264366997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,4096,0.015182221929232279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,4096,0.013399110900031196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,3584,0.011695110963450538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,3584,0.011119999819331698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,3584,0.012523555921183692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,3072,0.010876444478829702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,3072,0.010279110736317104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,3072,0.011563555234008364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,2560,0.00979911122057173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,2560,0.00907733374171787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,2560,0.01055466632048289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,2048,0.008443555898136562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,2048,0.008037333687146505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,2048,0.009400000174840292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,1536,0.006759111252095964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,1536,0.006908444480763541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,1536,0.00776977754301495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,1024,0.005836444596449534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,1024,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,1024,0.006861333217885759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,768,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,768,0.0041039999988343985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,768,0.006491555521885554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,512,0.005088888936572605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,512,0.0036728887094391715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,512,0.00611111107799742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,256,0.00464088883664873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,256,0.00328177772462368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,256,0.00573511090543535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3072,128,0.004362666772471534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,128,0.0030053332448005676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,128,0.005507555686765247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,64,0.0028524444335036804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,64,0.005798222290145026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,3072,32,0.0029519999192820657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,3072,32,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,65536,0.1380960014131334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,7168,0.02089599933889177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,65536,0.12350755267673069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,16384,0.035792890522215105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,16384,0.037295109695858426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,16384,0.036700444089041814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,1024,0.006011555592219035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,12288,0.02946044339074029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,12288,0.02863911125395033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,65536,0.12985866599612766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,12288,0.028502222564485338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,10240,0.02527911133236355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,10240,0.024832000335057575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,10240,0.02501511077086131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,8192,0.021147555775112573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,8192,0.020939555433061387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,8192,0.0206977774699529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,4096,2048,0.00902133352226681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,7168,0.019176888797018263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,7168,0.019377777973810833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,7168,0.01870044403605991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,6144,0.016864889197879367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,6144,0.01719466679626041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,6144,0.016911110944218107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,5120,0.014856000741322836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,5120,0.01311377767059538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,5120,0.014949333336618213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,4096,0.011136888629860349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,4096,0.012523555921183692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,4096,0.012792000340090858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,3584,0.011921777493423887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,3584,0.010279999838935005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,3072,0.0094035557574696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,3584,0.011744000017642975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,3072,0.01054666688044866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,3072,0.011045333411958484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,2560,0.009491555392742157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,2560,0.010034666293197209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,2560,0.008431110945012834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,2048,0.008292444050312042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,2048,0.008692444198661381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,2048,0.00740799970097012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,1536,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,1536,0.007537778053018782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,1024,0.005736888696750005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,1024,0.004809777769777509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,1024,0.006766222417354584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,768,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,768,0.004101333518822988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,768,0.0063457778758472866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,512,0.003669333126809862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,512,0.005112888912359874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,512,0.006068444500366847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,256,0.0032764443506797156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,256,0.0046364445653226645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,256,0.005762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,128,0.004392888810899523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,128,0.0029928889125585556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,128,0.005507555686765247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,64,0.002875555513633622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2560,32,0.0029208889851967492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,64,0.005724444571468566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2560,32,0.0058337777025169795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,65536,0.10093600220150417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,16384,0.03350399931271871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,65536,0.0995946658982171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,65536,0.1261404487821791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,16384,0.02981333269013299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,16384,0.02940622303220961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,12288,0.02667288813326094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,12288,0.024151111642519634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,10240,0.021222222182485793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,12288,0.022515555222829182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,10240,0.02313866714636485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,8192,0.0177839994430542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,10240,0.02054577734735277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,8192,0.017055999901559617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,7168,0.01737422247727712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,7168,0.016506666938463848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,7168,0.016180444094869826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,6144,0.015474667151769003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,6144,0.015062222878138224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,6144,0.014459555347760519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,5120,0.013388444152143268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,16384,0.04165600074662103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,5120,0.013415111435784234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,5120,0.01312355531586541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,4096,0.012098666694429187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,4096,0.011715555356608497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,4096,0.011501332951916588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,3584,0.010729778144094678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,3584,0.010576888918876648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,3584,0.01092177795039283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,3072,0.009851555857393477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,3072,0.009934222532643212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,3072,0.01000355515215132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,2560,0.007781333393520779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,2560,0.008943999807039896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,2048,0.007150222029950883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,2048,0.0070497774415545994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,2048,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,1536,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,1536,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,8192,0.023040889037979975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,1536,0.007010666860474481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,1024,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,1024,0.006521777974234686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,1024,0.0042151109211974675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,768,0.0039022221333450745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,768,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,768,0.006259555617968242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,512,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,512,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,256,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,256,0.004584889031118817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,256,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,128,0.003034666594531801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,128,0.004275555411974589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,128,0.0054560001525614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,64,0.0027866665687825945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,64,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,2048,32,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,2048,32,0.005648889061477449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,3584,5120,0.015771556231710646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,65536,0.0793093310462104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,65536,0.10338577959272598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,16384,0.02440266642305586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,65536,0.08937688668568929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,16384,0.025739555557568867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,12288,0.01945333348380195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,12288,0.02437155610985226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,12288,0.02126311096880171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,10240,0.017161778277821012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,10240,0.019275556008021038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,8192,0.018062222335073683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,8192,0.014451555079884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,8192,0.015599111715952555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,7168,0.016585777203241985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,7168,0.01349955548842748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,7168,0.01498933302031623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,6144,0.015012444721327888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,6144,0.012305777933862476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,6144,0.01387288918097814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,5120,0.013166222307417126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,5120,0.011261333194043903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,5120,0.012826667063766055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,4096,0.01073688848151101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,4096,0.010225777824719747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,4096,0.01072622256146537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,3584,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,3584,0.010016000105275048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,3584,0.010825778047243753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,3072,0.009340444372759925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,3072,0.009776888622177972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,3072,0.009335999687512716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,2560,0.007995555798212687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,2560,0.00756533361143536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,2560,0.008288000192907121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,2048,0.006695111178689533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,2048,0.006707555717892117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,2048,0.0075706665714581805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,1536,0.006025777922736273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,1536,0.0048942222363419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,1536,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,1024,0.005365333209435145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,1024,0.004192000048028098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,1024,0.006425777657164468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,768,0.0052328887912962176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,768,0.0038364442686239877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,768,0.006156444549560547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,512,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,512,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,512,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,256,0.004545777622196409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,256,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,256,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,128,0.004293333325121137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,128,0.0029297777348094513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,128,0.00545866663257281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,64,0.0028035555862718155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,64,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1536,32,0.0028880000528362063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1536,32,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,65536,0.10479022396935357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,65536,0.061140444543626576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,65536,0.07846933603286743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,16384,0.02881511052449544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,16384,0.018199111024538677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,16384,0.02386933399571313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,12288,0.02310666607485877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,12288,0.014705777168273926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,12288,0.019300444258583915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,10240,0.020124443703227572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,10240,0.01331999980741077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,10240,0.017831110292010836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,8192,0.011706666813956367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,8192,0.017277333471510146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,8192,0.015000889698664347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,7168,0.015463999576038785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,7168,0.010970667004585266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,7168,0.013914666242069669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,6144,0.013733333183659447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,6144,0.011961778004964193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,6144,0.01297688898113039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,5120,0.010567111273606619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,5120,0.012293332980738746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,5120,0.011503111157152386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,4096,0.00979555563794242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,4096,0.009191110730171204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,4096,0.009970666633711921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,3584,0.009155555731720394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,3584,0.008597333398130205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,3584,0.00924444446961085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,3072,0.007816889219813878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,3072,0.008611555728647444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,3072,0.008488888541857401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,2560,0.007088888850477006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,2560,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,2048,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,2048,0.006047111004590988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,2048,0.007344889144102733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,1536,0.006059555543793573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,1536,0.005049777941571342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,1536,0.0068488890926043195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,1024,0.005431111074156231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,1024,0.0041928887367248535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,1024,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,768,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,768,0.0038453332251972626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,768,0.006103111224042044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,512,0.004869333157936732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,512,0.003469333259595765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,512,0.005787555542257097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,256,0.004468444320890638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,256,0.003085333233078321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,256,0.005519111123349931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,128,0.004197333421972063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,128,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,128,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,64,0.0027093332674768236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,64,0.005397333453098933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,1024,32,0.0027520000520679685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,1024,32,0.005434666656785541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,65536,0.08547555738025242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,65536,0.05068177647060818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,65536,0.07306400272581312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,16384,0.027042667071024578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,16384,0.01516888870133294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,16384,0.022671111755900916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,12288,0.021367111139827307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,12288,0.012868444124857584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,12288,0.018329777651362948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,10240,0.01851733360025618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,10240,0.013923555612564087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,10240,0.01664355562792884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2560,1536,0.006810666786299811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,8192,0.016025778320100572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,8192,0.012093333734406365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,8192,0.014308444327778287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,7168,0.014510222607188754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,7168,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,7168,0.013407111167907715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,6144,0.013197333448463015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,6144,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,6144,0.012305777933862476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,5120,0.011260444091426002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,5120,0.009756444229020013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,5120,0.010785777535703448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,4096,0.00924800005224016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,4096,0.008821333448092142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,4096,0.009199110998047723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,3584,0.008803555534945594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,3584,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,3584,0.00868000007337994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,3072,0.008200888832410177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,3072,0.007190222541491191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,3072,0.008199111455016667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,2560,0.007439999944633908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,2560,0.006312888943486744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,2560,0.00777422222826216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,2048,0.0063662222690052455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,2048,0.005431111074156231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,8192,0.019222221440739103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,2048,0.007223111059930589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,1536,0.004770666774776247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,1536,0.0058337777025169795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,1536,0.006759111252095964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,1024,0.004038222134113312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,1024,0.0052977779673205475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,1024,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,768,0.005111111121045219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,768,0.003742222156789568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,768,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,512,0.004754666652944353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,512,0.0034231110993358824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,512,0.0057760001056724125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,256,0.004493333399295807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,2560,0.008581333690219456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,256,0.003039111072818438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,256,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,128,0.005392000079154968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,128,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,64,0.0053591111467944244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,64,0.0027804445061418745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,768,32,0.0027715555495686005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,768,32,0.005473777651786804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,65536,0.03866399990187751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,65536,0.06922755638758342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,16384,0.0210479994614919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,65536,0.06982044378916423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,16384,0.012349333200189801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,16384,0.022034666604465906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,12288,0.016677333248986136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,2048,512,0.005041777673694822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,12288,0.01792355544037289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,12288,0.011063111325105032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,10240,0.011194667054547204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,10240,0.016406221522225272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,8192,0.010094222095277574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,8192,0.012847111456924014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,8192,0.013905778527259827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,7168,0.011034666664070554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,7168,0.009575111170609793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,7168,0.012545777691735161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,6144,0.01014666673209932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,6144,0.00997866690158844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,6144,0.011252444651391772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,5120,0.009095999929640029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,5120,0.00927466650803884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,5120,0.010055999788973067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,4096,0.008066666622956594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,4096,0.007892444729804993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,4096,0.009054222040706212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,3584,0.00775999989774492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,3584,0.0075342224703894714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,3584,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,3072,0.007108444141017065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,3072,0.007237333390447829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,3072,0.008065777520338694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,2560,0.006188444379303191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,16384,0.03154666556252374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,2560,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,2048,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,2048,0.00608711110221015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,2048,0.007134222322040134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,1536,0.004651555584536659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,1536,0.006673777682913675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,1024,0.00508266646001074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,1024,0.004028444488843282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1536,10240,0.020762667059898376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,1024,0.006183111005359226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,768,0.0037457777394188773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,768,0.005994666781690385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,512,0.003359111025929451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,512,0.004657777647177379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,512,0.005718222094906702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,256,0.0030008889734745026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,256,0.005472888963090048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,128,0.0041653331783082755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,128,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,128,0.005340444544951121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,64,0.002704888789190186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,64,0.00535377777285046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,512,32,0.0026906666656335196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,512,32,0.005379555539952383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,65536,0.06949333349863689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,65536,0.0251875552866194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,65536,0.06792000267240736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,16384,0.020470221837361652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,16384,0.011239999698268043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,16384,0.021684444612926904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,12288,0.016202666693263583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,12288,0.009848888549539778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,12288,0.017111111018392775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,10240,0.013694222602579327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,10240,0.009121778110663096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,10240,0.015228443675571017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,8192,0.011768889096048145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,8192,0.008384000096056197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,8192,0.01276266657643848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,7168,0.010817777779367236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,1024,2560,0.007920888562997183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,7168,0.007870222131411234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,7168,0.01180266671710544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,6144,0.009870222045315636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,6144,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,6144,0.010848888920413123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,5120,0.008914666871229807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,5120,0.007085333267847697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,5120,0.009992000129487779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,4096,0.007878222399287755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,4096,0.006549333532651265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,4096,0.008940444224410588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,3584,0.006865777903132969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,3584,0.008544888761308458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,3072,0.007041778001520369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,3072,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,3072,0.00810222244924969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,2560,0.006550222221348021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,2560,0.005995555470387141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,2560,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,2048,0.006000888844331105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,2048,0.0053119998839166425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,2048,0.007065777977307637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,1536,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,1536,0.00463733325401942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,1536,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,1024,0.005088888936572605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,1024,0.003953777667548922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,1024,0.006152888966931238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,768,0.004839999808205498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,768,0.003622222277853224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,768,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,512,0.004554666578769684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,512,0.0033359999457995095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,512,0.005711110929648082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,256,0.004312889029582341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,256,0.002977777686383989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,256,0.005476444545719359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,128,0.004122666600677702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,128,0.0027679999669392905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,128,0.005316444569163852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,64,0.002623999905255106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,64,0.0053484443989064955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,256,32,0.002622222320901023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,256,32,0.0053582220441765254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,65536,0.06916177935070462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,65536,0.017328000730938382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,65536,0.06830755869547527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,16384,0.01954844428433312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,16384,0.008325333396593729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,16384,0.021032000581423443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,12288,0.015617777903874716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,12288,0.007535999847782983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,12288,0.016358221570650738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,10240,0.013607111242082385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,10240,0.0069982219073507525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,10240,0.014109333356221518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,8192,0.011711999773979187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,8192,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,8192,0.012128888732857175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,7168,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,7168,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,7168,0.011286222272449069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,6144,0.009759110709031424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,6144,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,6144,0.010471999645233154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,5120,0.008754666480753157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,5120,0.006718222051858902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,5120,0.009641778137948778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,4096,0.007820444802443186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,4096,0.0063528890411059065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,4096,0.00869688888390859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,3584,0.007343110938866933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,3584,0.006801777829726537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,3584,0.008213332957691616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,3072,0.006856888946559694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,3072,0.0064737776087390045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,768,128,0.00425866660144594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,3072,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,2560,0.006457777900828256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,2560,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,2560,0.007443555527263218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,2048,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,2048,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,2048,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,1536,0.005508444375462002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,1536,0.0046311111913787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,1536,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,1024,0.004951111144489712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,1024,0.003966222206751506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,1024,0.006089777582221561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,768,0.0047893333766195504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,768,0.003630222131808599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,768,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,512,0.004517333375083076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,512,0.0033271111961868075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,512,0.00565155554148886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,256,0.0029395555870400537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,256,0.005388444496525659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,128,0.004060444318585926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,128,0.002781333401799202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,128,0.005326222214433882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,64,0.0026364444444576898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,64,0.005327111317051782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,128,32,0.0026177778426143858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,128,32,0.005315555466545953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,65536,0.014089778065681458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,10240,0.014707555373509726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,16384,0.006692444284756978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,65536,0.06889333327611287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,16384,0.019322666856977675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,12288,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,12288,0.015574221809705099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,10240,0.006710222197903528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,10240,0.013908444179428948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,8192,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,7168,0.0064399999876817065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,8192,0.012071111136012606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,7168,0.011340444286664328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,6144,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,6144,0.01051999959680769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,5120,0.00666844430896971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,1536,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,5120,0.009722666607962714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,4096,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,3584,0.006813333266311222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,4096,0.008575999902354347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,3584,0.008169777691364288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,3072,0.006471111128727595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,2560,0.006647111227114995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,3072,0.007784888976150089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,2560,0.00740533322095871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,2560,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,2048,0.00686488880051507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,1536,0.004654222064548069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,2048,0.005345777918895085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,1536,0.006519111080302133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,256,0.004377777791685528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,1024,0.00612088872326745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,768,0.003585777762863371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,1024,0.003938666648334927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,768,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,512,768,0.004918222212129169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,512,0.005694222119119432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,256,0.0029164445069101122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,512,0.0032968889507982465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,256,0.00545866663257281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,128,0.0027946666296985415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,64,0.002603555512097147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,128,0.005320888840489917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,64,0.005340444544951121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,64,32,0.00259911103381051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,64,32,0.00535022219022115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,65536,0.0129466669427024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,16384,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,16384,0.006697777658700943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,10240,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,12288,0.006543111056089401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,65536,0.0677386654747857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,12288,0.015631111131774057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,8192,0.006536888993448681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,10240,0.013978666729397245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,7168,0.0063991112013657885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,8192,0.012010667059156628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,7168,0.011172444456153445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,6144,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,5120,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,6144,0.010437332921557956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,4096,0.006203555398517185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,4096,0.008453333543406593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,5120,0.009640889035330879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,3584,0.006767999794748094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,3584,0.008094222181373173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,3072,0.006414222220579783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,3072,0.007802666889296637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,2560,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,2560,0.0074311114019817775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,2048,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,1536,0.0046737777690092725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,2048,0.00685511115524504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,1024,0.003920888735188378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,1536,0.006496888895829518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,768,0.003632889025741153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,1024,0.006106666806671355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,768,0.0058400001790788435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,512,0.0032524443748924467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,256,0.002971555623743269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,512,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,256,0.005424889011515512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,128,0.002708444371819496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,128,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,64,0.0025759999536805684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,80,32,32,0.002567111204067866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,64,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,80,32,32,0.0053075556125905775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,12288,0.571306652492947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,16384,0.7399208810594348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,12288,0.29379023445977104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,16384,0.3844950993855794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,16384,0.38689955075581867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,10240,0.47343911064995664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,12288,0.29328444268968373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,10240,0.2646719879574246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,8192,0.37788799073961044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,8192,0.2131671110788981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,7168,0.3330542246500651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,7168,0.16380711396535239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,10240,0.24745244450039336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,8192,0.2014684412214491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,6144,0.285974211162991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,5120,0.23899645275539824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,6144,0.15062310960557726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,7168,0.18717777729034424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,6144,0.15206489298078749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,5120,0.1411502228842841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,4096,0.19111022684309217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,5120,0.1366693311267429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,4096,0.09908799992667304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,4096,0.10949866639243232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,3584,0.08699288633134629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,3584,0.17993244859907362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,3584,0.09114844269222683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,3072,0.14599200089772543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,2560,0.13093688752916124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,3072,0.08125778039296468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,2560,0.06536000304751925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,2048,0.10492444700664944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,2048,0.05401244428422716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,1536,0.08417689137988621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,1536,0.042615112331178456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,1536,0.04272177815437317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,1024,0.052128887838787503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,256,3584,0.007424888511498769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,1024,0.032215111785464816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,768,0.025391111771265667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,768,0.04236444499757555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,768,0.027170666389995154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,512,0.02012888921631707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,512,0.02985422147644891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,256,0.017445334129863314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,256,0.014858666393491956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,512,0.022188444932301838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,128,0.01144977741771274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,256,0.017435555656750996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,128,0.010255999863147736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,128,0.01465777721669939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,64,0.007903111477692923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,65536,32,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,64,0.01242844429281023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,65536,32,0.012092444631788464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,65536,0.7129617797003852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,16384,0.10947022173139785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,65536,0.41512356864081484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,16384,0.1896195544136895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,12288,0.07289777861701117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,16384,0.10877866877449884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,12288,0.13888800144195557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,10240,0.0616497794787089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,12288,0.09036178059048122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,10240,0.1153893338309394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,10240,0.06826844480302599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,8192,0.05049955513742235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,8192,0.09461777740054661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,8192,0.055897778934902616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,7168,0.045576890309651695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,7168,0.08518133560816447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,7168,0.05012977785534329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,6144,0.03999200132158067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,6144,0.07306133376227485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,6144,0.04393511017163595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,5120,0.03451910946104262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,5120,0.059881779882642955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,5120,0.03783822059631348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,4096,0.028782222006056044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,4096,0.05231377813551161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,4096,0.03136355678240458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,80,128,256,0.004266666869322459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,3584,0.04354844490687052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,3584,0.026907554931110803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,3584,0.028359111812379625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,3072,0.024028443627887305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,3072,0.038097778956095375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,3072,0.025918222135967676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,2560,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,2560,0.03511910968356662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,2560,0.022188444932301838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,2048,0.026702221896913316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,2048,0.01757866640885671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,2048,0.018908444378111098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,1536,0.01477777792347802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,1536,0.015930665863884818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,1536,0.020765332712067496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,1024,0.014666666587193808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,1024,0.011821333732869891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,1024,0.013083555632167392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,768,0.010431110858917236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,768,0.011552888486120435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,768,0.011739555332395764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,512,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,512,0.010109333528412713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,512,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,256,0.006133333262470033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,256,0.006537777682145436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,256,0.007229333122571309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,128,0.004885333279768626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,128,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,128,0.006774222271309958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,64,0.003638222399685118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,64,0.0069102222720781965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,16384,32,0.003742222156789568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,16384,32,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,2560,0.06581244203779432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,3072,0.07583466503355238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,2048,0.05567110909356011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,16384,0.13345511754353842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,65536,1024,0.033519112401538424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,65536,0.5565066867404515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,16384,0.0762044456270006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,12288,0.11988177564409043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,12288,0.05932355589336819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,12288,0.059918224811553955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,10240,0.09741866588592529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,10240,0.048688001102871366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,65536,0.29379023445977104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,8192,0.06928444570965238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,10240,0.05118755499521891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,8192,0.046922667158974536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,7168,0.036080890231662326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,7168,0.06100089020199246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,65536,0.29466221067640513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,7168,0.037596444288889565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,6144,0.03208266695340475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,6144,0.05379822187953525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,6144,0.033237334754731916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,5120,0.04497244291835361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,4096,0.03690133492151896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,5120,0.029083556599087183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,4096,0.023168888356950548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,4096,0.024311110377311707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,3584,0.0209840006298489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,3584,0.032808889945348106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,3584,0.022184888521830242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,3072,0.01870044403605991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,3072,0.028778665595584448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,3072,0.019926221834288705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,2560,0.024558222956127588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,2560,0.01759199963675605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,2560,0.016875555117925007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,2048,0.0203840004073249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,2048,0.014548444085650973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,1536,0.016116445263226826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,2048,0.015248888068728976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,1536,0.012281777958075205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,1536,0.01313244468635983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,1024,0.011342222491900126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,1024,0.010081777969996134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,1024,0.01130844404300054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,768,0.009111999637550777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,768,0.009360888765917884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,768,0.0101724440852801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,512,0.007270221908887227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,512,0.007874666816658443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,512,0.008203555312421586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,256,0.004412444515360726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,256,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,256,0.006486222147941589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,128,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,128,0.006081777728266186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,64,0.0033235556135574975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,64,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,12288,32,0.00638488887084855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,12288,32,0.0034551111360390982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,65536,0.46363019943237305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,16384,65536,0.35566311412387425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,16384,0.07170844740337796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,16384,0.12152622805701362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,65536,0.24314400884840223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,65536,0.25143734614054364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,16384,0.06722399923536512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,12288,0.09479377667109172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,10240,0.04228622383541531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,12288,0.053151110808054604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,10240,0.07161066929499309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,10240,0.045090665419896446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,8192,0.035170667701297335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,8192,0.05803200271394518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,8192,0.03716622127426995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,7168,0.031488888793521456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,7168,0.05132088727421231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,7168,0.03340977761480544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,6144,0.027638221780459087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,6144,0.04534577661090427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,6144,0.02978044417169359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,5120,0.02390755547417535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,5120,0.038112001286612615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,5120,0.02604977786540985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,4096,0.02015644477473365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,4096,0.031395554542541504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,4096,0.021917333205540974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,3584,0.02787289023399353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,3584,0.020076443751653034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,3072,0.016736000776290894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,3072,0.024587555064095393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,3072,0.01805688937505086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,2560,0.015049777097172208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,2560,0.021126222279336717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,2560,0.016183111402723525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,2048,0.013024888932704926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,2048,0.017443555924627516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,1536,0.011316444310877057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,1536,0.013740444348918067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,2048,0.014350222216712104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,1536,0.012448888685968188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,1024,0.009926222264766693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,1024,0.0094355551732911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,1024,0.010747555229398938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,768,0.008581333690219456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,768,0.008311111066076491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,768,0.009909333454238044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,512,0.0069066666894488865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,512,0.006559999866618051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,512,0.007727999654081132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,256,0.005052444421582752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,256,0.003834666477309333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,256,0.00646577775478363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,128,0.0047422221137417685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,128,0.0034888887570963967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,128,0.006141333530346553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,64,0.0031555555760860443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,64,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,10240,32,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,10240,32,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,65536,0.36580443382263184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,65536,0.19704088899824354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,16384,0.051983998881445996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,16384,0.07422844568888347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,16384,0.0906204448805915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,16384,0.06112622552447849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,65536,0.24466310607062447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,12288,0.04124177826775445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,12288,0.06925511360168457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,10240,0.03496355480617947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,10240,0.058111111323038735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,12288,0.0503475566705068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,8192,0.04056444433000352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,10240,0.0417795545525021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,8192,0.04731377628114489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,8192,0.03464266657829285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,7168,0.04200177722507053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,7168,0.025967111190160114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,5120,0.027520888381534155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,7168,0.031131555636723835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,6144,0.023112000690566167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,6144,0.03693600164519416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,5120,0.0313297775056627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,5120,0.02442399991883172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,6144,0.02756622268093957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,4096,0.017286222842004564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,4096,0.02946399980121189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,4096,0.020402666595247056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,3584,0.015959110524919298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,3584,0.023013333479563396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,3584,0.018418666389253404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,3072,0.02049422264099121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,3072,0.014468444718254937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,3072,0.01677688956260681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,2560,0.01752622259987725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,12288,128,0.0047235555118984645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,2560,0.012984889249006907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,2560,0.01516355574131012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,2048,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,2048,0.013212444053755866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,2048,0.014506666196717156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,1536,0.011550222006109027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,1536,0.009943111075295342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,1536,0.011863110793961419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,1024,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,1024,0.010114666488435533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,768,0.007745777567227681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,768,0.008605333666006723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,768,0.007375111182530721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,512,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,512,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,512,0.006806222101052602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,256,0.004771555463473002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,256,0.0036764442920684814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,256,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,128,0.0034302222645945023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,128,0.006167999986145232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,64,0.0032755554550223877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,64,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,8192,32,0.0033928888539473214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,8192,32,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,65536,0.17694311671786836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,12288,0.049398223559061684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,65536,0.3368071185217963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,16384,0.0533955560790168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,16384,0.08624088764190674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,16384,0.04852622085147434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,12288,0.037969777981440224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,12288,0.0613831082979838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,65536,0.18268088499704996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,12288,0.03884977764553494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,10240,0.03206222256024679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,10240,0.059155556890699595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,10240,0.033492445945739746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,8192,0.02677511175473531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,8192,0.043281777037514575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,7168,0.03707999984423319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,8192,0.027856889698240492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,7168,0.024160888459947374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,10240,3584,0.018592889110247295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,7168,0.02529955572552151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,6144,0.033634665939542986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,6144,0.021723555194007024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,5120,0.028646220763524372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,6144,0.022878222995334204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,5120,0.02015555567211575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,4096,0.016414221790101793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,4096,0.024338665935728285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,3584,0.021542222963439092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,4096,0.01696977847152286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,3584,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,3584,0.015740444262822468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,3072,0.018311111463440787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,3072,0.013823111024167804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,3072,0.014431110686726041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,2560,0.012283555335468717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,2560,0.0163137780295478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,2560,0.013196444345845116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,2048,0.013474666410022311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,2048,0.010858666565683154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,1536,0.010874666273593903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,2048,0.011786667009194693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,1536,0.009552000297440423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,1024,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,1536,0.010438222024175854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,1024,0.008170666793982187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,768,0.007089777953094906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,1024,0.00906488878859414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,768,0.007072000039948358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,768,0.006741333338949416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,512,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,512,0.005140444470776452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,512,0.006171555568774541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,256,0.0047217777205838096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,256,0.0035084444615576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,256,0.005774222314357758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,128,0.004469333423508538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,128,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,128,0.0033893332713180115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,64,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,32,0.003173333282272021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,7168,64,0.003152888889114062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,7168,32,0.005671999934646819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,65536,0.2762400044335259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,16384,0.04639200038380093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,65536,0.14860444598727757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,65536,0.15736177232530382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,16384,0.07447644737031725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,16384,0.04381510946485731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,12288,0.03456266721089681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,12288,0.05860444572236803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,10240,0.046625776423348315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,12288,0.03497511148452759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,10240,0.028063111835055884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,10240,0.03051111102104187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,8192,0.038217776351504855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,8192,0.02355466617478265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,7168,0.03388088941574097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,7168,0.02127200033929613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,8192,0.025598223010698955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,6144,0.029987557066811457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,7168,0.023341332872708637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,6144,0.018746667438083224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,5120,0.02478222217824724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,5120,0.01685688893000285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,6144,0.021038222644064162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,5120,0.01868799991077847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,4096,0.014519110321998596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,4096,0.020436444216304354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,4096,0.015847111741701763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,3584,0.018215111560291715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,3584,0.013565333353148567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,3584,0.014741332994567024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,3072,0.012302222351233164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,3072,0.016265778077973258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,3072,0.013704000247849358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,2560,0.011390222443474663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,2560,0.014071111877759298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,2560,0.012501333322789935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,2048,0.009936889012654623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,2048,0.011880888707107969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,2048,0.011119999819331698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,1536,0.008997333546479544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,1536,0.009744000103738572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,1536,0.00998399986161126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,1024,0.007778666913509369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,1024,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,1024,0.007806222471925948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,768,0.006542222367392645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,768,0.005705777969625261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,768,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,512,0.005114666703674528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,512,0.004736888739797804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,512,0.006083555519580841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,256,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,256,0.004743111216359668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,256,0.005762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,128,0.003277333246337043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,6144,128,0.004567111117972268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,128,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,64,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,64,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,6144,32,0.003022222262289789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,6144,32,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,8192,0.029061334000693426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,5120,0.020095111595259774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,65536,0.23097332318623862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,16384,0.07091377841101752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,65536,0.1295457813474867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,1024,0.008573333422342936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,16384,0.037645333343082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,65536,0.1276551087697347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,12288,0.029543999168607924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,12288,0.030371556679407757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,10240,0.03844889005025228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,12288,0.049785776270760425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,8192,128,0.004518222063779831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,10240,0.02571733295917511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,8192,0.03148266673088074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,10240,0.026477333572175767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,8192,0.021486222743988037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,8192,0.022247110803922016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,7168,0.028111110130945843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,7168,0.01943644384543101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,7168,0.020371556282043457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,6144,0.017768000562985737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,5120,0.01588622232278188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,6144,0.018461333380805123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,5120,0.02123022245036231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,6144,0.02514044443766276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,5120,0.01662133302953508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,4096,0.01777688827779558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,4096,0.014105778601434497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,3584,0.013239999612172445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,3584,0.01587199999226464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,4096,0.013587555951542325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,3072,0.013879111243618859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,3072,0.01236533373594284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,2560,0.012064889073371887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,2560,0.0105262224872907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,2560,0.011471110913488599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,2048,0.010213333699438307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,2048,0.009406222237481011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,2048,0.01033688916100396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,1536,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,1536,0.008460444708665213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,1536,0.009371555513805812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,1024,0.006179555422729916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,1024,0.006806222101052602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,1024,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,768,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,768,0.006165333506133821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,768,0.005330666485759947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,512,0.0041031113101376426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,512,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,256,0.004664888812435998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,256,0.0033599999215867785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,256,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,128,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,128,0.00546577779783143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,128,0.004451555510361989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,64,0.0029724445194005966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,64,0.005496888938877318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,5120,32,0.003015999992688497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,5120,32,0.005504000104135937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,65536,0.18485332859887016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,65536,0.1034453312555949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,16384,0.03074044320318434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,16384,0.05510222249560886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,65536,0.11495378282335068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,16384,0.03215644425816006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,12288,0.024358222881952923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,12288,0.04089955488840739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,12288,0.026523555318514507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,10240,0.02104355560408698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,10240,0.03498133354716831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,10240,0.02314222190115187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,8192,0.018000000052981906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,8192,0.02935111191537645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,8192,0.01940088967482249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,7168,0.016491555505328707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,7168,0.026663111315833196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,7168,0.01794933279355367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,6144,0.014961777461899651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,6144,0.021104888783560857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,6144,0.016315556234783597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,5120,0.013403555585278405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,5120,0.01808177762561374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,5120,0.014764444695578681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,4096,0.011785777906576792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,4096,0.01477866702609592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,4096,0.012710221939616732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,3584,0.011380444798204633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,3584,0.01328888866636488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,3584,0.011935999823941125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,3072,0.010187555518415239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,3072,0.011964444484975604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,3072,0.011160888605647616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,2560,0.009553777674833933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,2560,0.010455110834704505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,2560,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,2048,0.009195555415418413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,2048,0.00853777759604984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,2048,0.00961866643693712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,1536,0.007808888951937358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,1536,0.00758755538198683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,7168,5120,0.018816888332366943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,1536,0.008367111285527548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,1024,0.0053582220441765254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,1024,0.00609777785009808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,1024,0.0064080001579390625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,768,0.00507733349998792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,768,0.004819555415047539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,768,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,512,0.005780444376998478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,512,0.004820444517665439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,512,0.003680888977315691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,256,0.0032088889016045462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,256,0.004554666578769684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,256,0.005518222020732031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,4096,128,0.004335999902751711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,128,0.0029688889367712867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,64,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,128,0.005427555491526921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,64,0.005436444448100195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,4096,32,0.002858666703104973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,4096,32,0.005427555491526921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,16384,0.02853333287768894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,65536,0.16210844781663683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,16384,0.048206223381890186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,12288,0.022648889157507155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,65536,0.09652622540791829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,16384,0.03315022256639268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,65536,0.09839110904269749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,12288,0.03794577717781067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,10240,0.019684443871180218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,12288,0.02474044428931342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,10240,0.03240088952912225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,10240,0.021817778547604878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,8192,0.017047999633683097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,8192,0.027727999620967444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,8192,0.018079999420377944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,7168,0.015458666616015963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,7168,0.016907556189431083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,7168,0.024941333466105994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,6144,0.01420088940196567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,6144,0.02232355541653103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,5120,0.016145777371194627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,6144,0.015487111277050443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,5120,0.014072888427310519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,4096,0.013240888714790344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,4096,0.011401777466138205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,4096,0.012141333685980903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,3584,0.01205599970287747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,3584,0.010543111297819348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,3072,0.01090133355723487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,3584,0.011414222419261932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,3072,0.009956444303194681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,3072,0.010778666370444827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,2560,0.009776888622177972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,2560,0.009202666580677032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,2560,0.010129777921570672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,2048,0.008257778154479133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,2048,0.009296000003814697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,2048,0.009010666774378883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,1536,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,1536,0.00740533322095871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,1536,0.007115555306275685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,1024,0.0053688887920644544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,1024,0.0058337777025169795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,1024,0.00619911112719112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,768,0.005046222358942032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,768,0.004256000121434529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,768,0.005982222242487802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,512,0.004863111095296012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,512,0.0036524443162812125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,512,0.005819555371999741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,256,0.004474666797452503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,256,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,256,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,128,0.004316444612211651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,128,0.002969777832428614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,128,0.005414222263627582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,64,0.002858666703104973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3584,32,0.0028728888266616394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,64,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3584,32,0.005428444594144821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,65536,0.14315644900004068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,65536,0.08549155791600545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,16384,0.0248391115003162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,16384,0.04175733195410835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,65536,0.09126310878329807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,16384,0.027820444769329492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,12288,0.03281955586539374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,12288,0.022802667485343084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,10240,0.01791200041770935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,10240,0.02823466724819607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,10240,0.02013422217633989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,8192,0.015948444604873657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,8192,0.024580443898836773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,8192,0.016858667135238647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,7168,0.014627556006113688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,7168,0.021985777550273474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,7168,0.015832000308566622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,6144,0.013266666895813413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,6144,0.01704088846842448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,6144,0.014567111929257711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,5120,0.01219466659757826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,5120,0.014504000544548035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,5120,0.013420444395807056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,4096,0.010954666468832227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,16384,0.03749955693880717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,4096,0.012056888805495368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,4096,0.011591999895042844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,3584,0.010250666903124915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,3584,0.011039111349317761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,3584,0.011141333315107556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,3072,0.010015111002657149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,3072,0.00960888879166709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,3072,0.010447999669445885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,2560,0.008880889250172509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,2560,0.008919999831252629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,2560,0.009747555686367882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,2048,0.007886221839321984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,2048,0.008400888906584846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,1536,0.006098666538794835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,1536,0.006830222076839871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,1536,0.007034666836261749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,1024,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,1024,0.005254222287072076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,1024,0.006180444525347815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,768,0.0041964443193541635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,768,0.005995555470387141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,512,0.004780444420046276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,512,0.003642666671011183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,512,0.005780444376998478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,256,0.004456000195609199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,256,0.0032133333798911837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,256,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,3584,0.012630222572220696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,128,0.0029200000895394217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,128,0.004239999999602635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,64,0.0028168888141711554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,128,0.005423999908897612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,64,0.005380444642570283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,3072,32,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,3072,32,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,3072,0.011691555380821228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,5120,512,0.005013333426581489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,16384,0.02296977738539378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,65536,0.1217457718319363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,16384,0.03486311104562547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,16384,0.02606666584809621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,65536,0.07507733503977458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,12288,0.01866666641500261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,65536,0.08223555485407512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,12288,0.027962668074501887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,12288,0.021368889345063105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,10240,0.0165928883685006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,10240,0.018944000204404194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,10240,0.02440533373090956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,8192,0.014519999424616495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,8192,0.020743111769358318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,8192,0.01603200038274129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,7168,0.013425778183672162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,7168,0.018811555372344125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,7168,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,6144,0.015118221441904703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,6144,0.014019555515713163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,5120,0.011494222614500256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,5120,0.01293688929743237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,5120,0.012871111432711283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,4096,0.01037688884470198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,4096,0.010905777414639791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,4096,0.011164444188276926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,3584,0.009648000200589498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,3584,0.010024888647927178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,3584,0.01070844464831882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,3072,0.009111999637550777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,3072,0.00925333301226298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,3072,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,2560,0.008471999731328752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,2560,0.008247111406591203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,2560,0.009368889033794403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,2048,0.006715555571847492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,2048,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,2048,0.00756444450881746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,1536,0.005821333163314396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,1536,0.0064479998416370815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,1536,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,1024,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,1024,0.004654222064548069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,1024,0.006170666466156642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,768,0.005037333402368757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,768,0.003983111017280155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,768,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,512,0.003600888782077365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,512,0.005778666585683823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,256,0.004443555656406614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,256,0.0032319999817344877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,256,0.005518222020732031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,128,0.0041964443193541635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,128,0.002945777856641345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,128,0.005355555564165115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,64,0.0027991111079851785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,64,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2560,32,0.0028595555987623003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2560,32,0.005387555393907759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,65536,0.07040177451239692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,65536,0.09923289219538371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,65536,0.07620888948440552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,16384,0.02157955533928341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,16384,0.029166221618652344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,16384,0.024474667178259954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,12288,0.017431111799346078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,12288,0.02356622285313076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,12288,0.02001688877741496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,10240,0.015431111057599386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,10240,0.020672889219390023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,10240,0.017853332890404593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,8192,0.01330044451687071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,8192,0.017257778180970084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,8192,0.015146666102939181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,7168,0.01591111057334476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,7168,0.014275555809338888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,6144,0.01148000028398302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,6144,0.01313244468635983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,6144,0.013359111216333179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,5120,0.010775999890433418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,5120,0.012697777814335294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,5120,0.012311999996503195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,4096,0.009474666582213508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,4096,0.011252444651391772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3584,5120,0.012815111213260226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,4096,0.010665777656767102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,3584,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,3584,0.01015377789735794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,3072,0.008159110943476358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,3072,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,3072,0.009349333743254343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,2560,0.0070017774899800616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,2560,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,2560,0.00776088900036282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,2048,0.006063111126422882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,2048,0.0068915556702348925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,2048,0.006816000160243776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,1536,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,1536,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,1536,0.0064497776329517365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,1024,0.005135111096832487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,1024,0.004110222061475118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,1024,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,768,0.004942222187916438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,768,0.0038035557501845886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,768,0.0058551111982928375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,512,0.004709333181381226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,512,0.0034773333205117118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,512,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,256,0.004399999976158142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,256,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,256,0.005396444350481033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,128,0.004265777766704559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,128,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,128,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,64,0.0027502222607533136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,12288,0.020265777905782063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,64,0.005355555564165115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,2048,32,0.002785777880085839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,2048,32,0.0053288886944452924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,65536,0.06945155726538764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,65536,0.07767377959357367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,65536,0.07115555471844144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,2048,0.007511110769377814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,16384,0.023916444844669763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,16384,0.022816000713242426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,12288,0.016726222303178575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,3072,768,0.005063111169470681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,12288,0.019152889649073284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,12288,0.018947554959191214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,10240,0.014886221951908536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,10240,0.01678311162524753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,10240,0.016934222645229764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,8192,0.013040000365840064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,8192,0.014308444327778287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,8192,0.014707555373509726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,7168,0.013412444127930535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,7168,0.013714666995737286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,6144,0.011215111447705163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,6144,0.012128888732857175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,6144,0.012850667039553324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,5120,0.01017155581050449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,5120,0.011232000258233813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,5120,0.011904888682895236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,4096,0.009169778062237633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,4096,0.010272888673676385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,4096,0.010208889014191097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,3584,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,3584,0.008461332983440822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,3584,0.009468444519572789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,3072,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,3072,0.007353777686754863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,3072,0.007848000360859765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,2560,0.006719999843173557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,2560,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,2560,0.0073928890956772705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,2048,0.006040888941950268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,2048,0.006401777681377199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,2048,0.006853333363930385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,1536,0.00565155554148886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,1536,0.004831999954250124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,1536,0.006514666808976068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,1024,0.005130666825506422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,1024,0.004118222329351637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,1024,0.006065777606434292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,768,0.004896000027656555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,768,0.0037519998020595978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,768,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,512,0.004695999953481886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,512,0.0034426665968365143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,512,0.005679111099905438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,256,0.00433955548538102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,256,0.003096888876623578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,256,0.005495111147562663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,128,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,128,0.0053635554181204895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,64,0.002712888850106133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,64,0.005378666851255629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1536,32,0.002752888947725296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1536,32,0.005392888767851724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,6144,0.012514666550689273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,65536,0.0556817783249749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,65536,0.06954222255282931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,16384,0.020959110723601446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,65536,0.06813155280219184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,16384,0.017982222967677645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,16384,0.021708443760871887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,12288,0.01683822274208069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,12288,0.014575110541449653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,12288,0.018124444617165458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,10240,0.014701333310869006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,10240,0.013181333740552267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,10240,0.01626133256488376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2560,512,0.0047457776963710785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,8192,0.012859555582205454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,8192,0.011233777635627322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,8192,0.01405244403415256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,7168,0.010637332995732626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,7168,0.011901333100265928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,7168,0.013264888690577613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,6144,0.010703999963071613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,6144,0.012225777738624148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,5120,0.009393778112199571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,5120,0.009767110976907942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,5120,0.011046222514576383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,4096,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,4096,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,4096,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,3584,0.007520889242490132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,3584,0.00870666652917862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,3584,0.008398222426573435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,3072,0.00700711127784517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,3072,0.008037333687146505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,3072,0.007703999678293864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,2560,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,2560,0.007262222468852997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,7168,0.012578667038016848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,2048,0.006063999815119638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,2048,0.00564444437623024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,2048,0.006883555402358373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,1536,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,1536,0.004821333206362194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,1536,0.006504888749784893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,1024,0.004117333226733738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,1024,0.005095110999213325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,1024,0.006075555665625467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,768,0.00488711107108328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,768,0.0037564444873068067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,768,0.005856888989607493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,512,0.0034551111360390982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,512,0.005647999958859549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,256,0.00434666665063964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,256,0.0030453333424197305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,256,0.005446222093370225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,128,0.004137777619891697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,128,0.0028080000645584534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,128,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,64,0.0029128889242808023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,2048,3584,0.009081777599122789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,64,0.005335999859703912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,1024,32,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,1024,32,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,65536,0.04541688826349047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,65536,0.06892177793714735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,16384,0.020706666840447318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,65536,0.06736622254053752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,16384,0.01500888831085629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,16384,0.021614222062958613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,12288,0.016551110479566786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,12288,0.012726222475369772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,12288,0.017854221993022494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,10240,0.011760888828171624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,10240,0.016227554943826463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,8192,0.012652444342772165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,8192,0.010457778142558204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,8192,0.013920888304710388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,7168,0.01143022212717268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,7168,0.009932444327407414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,7168,0.012637333737479316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,6144,0.010065777434243096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,6144,0.009313777916961247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,6144,0.01127644462717904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,5120,0.008935111264387766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,5120,0.008748444418112436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,5120,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,4096,0.007874666816658443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,4096,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,4096,0.00847466703918245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,3584,0.007449777589903937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,3584,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,3584,0.008095111283991072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,3072,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,3072,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,3072,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,2560,0.0064737776087390045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,2560,0.006100444330109491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,2560,0.007296000089910295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,2048,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,2048,0.005364444520738389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,2048,0.006844444407357111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,16384,0.021074665917290583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,1536,0.0047155556579430895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,1536,0.006474666711356904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,1024,0.005058666484223472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,1024,0.004064889003833135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,1024,0.00609422226746877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,768,0.004837333328194088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,768,0.0037066667444176147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,768,0.005868444426192178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,512,0.004602666530344221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,512,0.0033928888539473214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,512,0.005689777847793367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,256,0.004311111238267687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,256,0.0030257778449190986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,256,0.005483555710977978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,128,0.004126222183307012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,128,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,128,0.005362666729423735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,64,0.002704000100493431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,64,0.0053582220441765254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,768,32,0.0027111110587914786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,768,32,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,7168,0.012039110892348819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,65536,0.03408711155255636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,65536,0.06884355677498712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,16384,0.020469332734743755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,16384,0.012206222448084088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,65536,0.06667288806703356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,16384,0.021532444490326777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,12288,0.010728889041476779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,12288,0.01773244473669264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,10240,0.014410666293568082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,10240,0.009767110976907942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,10240,0.015973332855436537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,8192,0.011951111257076263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,8192,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1536,128,0.0041777777175108595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,8192,0.01277511070171992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,7168,0.010785777535703448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,7168,0.0087333329849773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,7168,0.011550222006109027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,6144,0.009740444521109263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,6144,0.007900444169839224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,6144,0.010327999790509542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,5120,0.008813333180215623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,5120,0.007727999654081132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,5120,0.009529777699046666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,4096,0.007877333296669854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,4096,0.006691555596060223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,4096,0.008456000023418004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,3584,0.007418666448858049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,3584,0.007207999626795451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,3584,0.008091555701361762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,3072,0.006882666713661618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,3072,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,3072,0.007713778151406183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,6144,0.010878221856223213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,2560,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,2560,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,2048,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,2048,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,2048,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,2560,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,1536,0.0046746668716271715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,1536,0.006452444526884291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,1024,0.004981333182917701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,1024,0.003990222182538775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,1024,0.006043555421961679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,768,0.004741333425045013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,768,0.003649777836269803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,768,0.0058373332851462895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,512,0.004552000098758274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,512,0.0033173333439562055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,512,0.005631999837027655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,256,0.004224888980388641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,256,0.0029519999192820657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,256,0.005415110952324337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,128,0.004087999877002505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,128,0.0027679999669392905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,128,0.005295111073387994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,64,0.0026275554878844153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,64,0.0053119998839166425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,512,32,0.002624888800912433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,512,32,0.0053182223604785065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,65536,0.02353688908947839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,65536,0.06661333640416463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,16384,0.019896888070636325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,16384,0.009328000247478485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,16384,0.020870221985710993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,1024,512,0.0046773333516385825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,12288,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,12288,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,10240,0.014681778020328946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,10240,0.00776444458299213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,10240,0.014483556151390076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,768,1536,0.005533333453867171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,8192,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,8192,0.011972444752852121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,7168,0.010703999963071613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,7168,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,7168,0.011184888581434885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,6144,0.009806222385830348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,6144,0.006714666469229593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,6144,0.01036800030204985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,5120,0.00886933339966668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,5120,0.0068764446510208984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,5120,0.009546666509575313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,4096,0.007882666256692674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,4096,0.006477333191368315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,4096,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,3584,0.0073928890956772705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,3584,0.006909333169460297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,3584,0.0081395556529363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,3072,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,3072,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,3072,0.007712889048788283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,2560,0.006439111298984951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,2560,0.006020444548792309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,2560,0.007325333025720384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,2048,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,2048,0.0053546664615472155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,2048,0.006862222320503659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,1536,0.005433777968088786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,1536,0.0046942221621672315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,1536,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,1024,0.004970666848950916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,1024,0.003984000119898054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,1024,0.00608800020482805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,768,0.004713777866628435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,768,0.00361333332127995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,768,0.005819555371999741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,512,0.004457777573002709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,512,0.003318222239613533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,512,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,256,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,256,0.002975111206372579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,12288,0.016406221522225272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,256,0.005435555759403441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,2560,0.006498666687144174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,128,0.0027884443600972495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,128,0.005294222384691238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,64,0.0026266665922270883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,64,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,256,32,0.0026382222357723448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,256,32,0.0052871112194326185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,65536,0.016538666354285348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,65536,0.06660088565614489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,512,1536,0.005445333404673471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,16384,0.019208888212839764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,16384,0.007521777517265744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,16384,0.02028622229894002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,12288,0.015233778291278414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,12288,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,12288,0.015551111764378019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,10240,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,10240,0.013438222308953604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,10240,0.013572444518407186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,8192,0.011583111352390714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,8192,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,8192,0.011854222251309289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,7168,0.006477333191368315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,7168,0.011036444041464063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,6144,0.009667555491129557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,6144,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,6144,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,5120,0.008719999757077958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,5120,0.006695999867386288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,5120,0.009489778015348647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,4096,0.007788444558779399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,4096,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,4096,0.008451555338170793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,3584,0.007342221836249034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,3584,0.006759111252095964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,3584,0.008059555457697975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,3072,0.006804444309737947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,3072,0.0064284445510970215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,3072,0.007644444704055786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,2560,0.0063662222690052455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,2560,0.00599911105301645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,2560,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,2048,0.005807999935415056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,2048,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,2048,0.006808888994985157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,1536,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,1536,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,1536,0.006430222342411677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,1024,0.004921777794758479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,1024,0.003964444415436851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,1024,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,768,0.0046826667255825466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,768,0.003638222399685118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,768,0.0058240000572469496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,512,0.004440000073777305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,512,0.003321777822242843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,512,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,65536,0.0684551133049859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,256,0.002924444360865487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,256,0.0041911109454101985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,256,0.005415110952324337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,128,0.0027831111931138565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,128,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,64,0.002605333303411802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,64,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,128,32,0.0025964443468385274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,128,32,0.0052977779673205475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,65536,0.014665777484575907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,65536,0.06662933031717937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,16384,0.006545777950021956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,16384,0.018943111101786297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,12288,0.006383111079533895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,12288,0.015284443895022074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,10240,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,10240,0.01370488852262497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,8192,0.0064791109826829694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,8192,0.011902222202883827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,7168,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,7168,0.01108888867828581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,6144,0.006156444549560547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,6144,0.010306666294733683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,5120,0.006522666662931442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,12288,0.015486222174432544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,5120,0.009483555952707926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,4096,0.00610133343272739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,4096,0.008450667063395182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,3584,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,3584,0.008103999826643202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,3072,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,3072,0.007700444095664554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,2560,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,2560,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,2048,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,2048,0.006826666494210561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,1536,0.004610666798220741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,1536,0.006447111152940326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,1024,0.003937777959638172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,768,0.003559111307064692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,1024,0.006009777800904379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,768,0.0058471109304163195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,512,0.0032542221662071017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,512,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,256,0.0029404444826973807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,256,0.005399111244413588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,128,0.0027457777824666765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,128,0.0052968888647026485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,64,0.0026088888860411118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,64,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,64,32,0.0025884444928831523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,64,32,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,65536,0.013393777940008374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,65536,0.06658044126298693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,16384,0.006471111128727595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,16384,0.018917333748605516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,12288,0.006411555740568373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,12288,0.015244444211324057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,10240,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,10240,0.013660444153679742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,8192,0.0064328888224230866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,8192,0.011768889096048145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,10240,0.013602666556835175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,7168,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,6144,0.006100444330109491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,7168,0.011071999867757162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,6144,0.010168000227875179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,5120,0.006511111226346757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,5120,0.009379555781682333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,4096,0.006099555641412735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,4096,0.008381333616044786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,3584,0.006488889041874144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,3584,0.008014221986134848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,3072,0.006146666490369373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,3072,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,2560,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,2560,0.007186666958861881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,2048,0.0052533331844541765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,2048,0.006735111276308696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,1536,0.004616888860861461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,1024,0.003936888857020272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,1536,0.006340444501903322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,1024,0.006032889087994893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,768,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,768,0.005838222387764189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,512,0.0032435556252797446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,512,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,256,0.002887111157178879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,256,0.005421333428886201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,128,0.0027315556589100095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,64,0.002571555475393931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,128,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,64,32,32,0.002612444468670421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,64,0.005223111146026188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,64,32,32,0.0052595556610160405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,8192,0.011570666399266986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,16384,0.7183635499742297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,16384,0.3794995678795709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,12288,0.5462480121188694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,12288,0.31535644001430935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,16384,0.38658666610717773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,10240,0.4561653137207031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,12288,0.27613332536485463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,10240,0.2642933262719048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,8192,0.36635467741224503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,256,128,0.0040942223535643685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,8192,0.20972088972727457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,10240,0.24643556276957193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,8192,0.20963999960157606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,7168,0.3184071116977268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,7168,0.16033421622382268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,6144,0.2751031186845568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,6144,0.16192977958255345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,7168,0.17163466082678902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,5120,0.2290711137983534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,5120,0.12425777647230361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,4096,0.18664177258809408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,6144,0.16224443912506104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,3584,0.1630257765452067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,5120,0.13094488779703775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,4096,0.10137155320909287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,3584,0.08598933617273967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,3072,0.07463466458850436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,3072,0.14133777883317736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,3584,0.08909510903888279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,3072,0.07612711191177368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,2560,0.11828000015682644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,2560,0.06439289119508532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,2560,0.06497333447138469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,2048,0.09523822201622857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,2048,0.0532551109790802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,1536,0.0736044446627299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,2048,0.053694221708509654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,1536,0.0421982208887736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,1024,0.03044711218939887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,1024,0.05051911208364698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,65536,0.06843466891182794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,1024,0.03173866536882188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,768,0.04045688774850633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,768,0.024780443973011438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,768,0.026759111218982275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,512,0.028674665424558852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,512,0.01958311100800832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,256,0.016304888659053378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,256,0.014494222071435718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,512,0.021662222014533147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,256,0.016975111431545682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,128,0.011147555377748279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,128,0.00981600003110038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,128,0.014291556345091926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,64,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,64,0.011623111036088733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,65536,32,0.005671999934646819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,65536,32,0.011610666910807291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,7168,0.010564444793595208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,65536,0.7094995710584852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,16384,0.2077724403805203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,65536,0.3784622351328532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,65536,0.4148915608723958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,16384,0.10538666778140598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,12288,0.1434017817179362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,12288,0.07692977454927233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,64,128,128,0.004025777594910728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,12288,0.07995822032292683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,10240,0.11940088537004258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,10240,0.061405334207746715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,8192,0.09072799815071954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,10240,0.07088889016045465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,8192,0.05063466562165154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,7168,0.07981244723002116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,8192,0.05749422311782837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,7168,0.04530133472548591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,6144,0.03973689013057285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,6144,0.0698417756292555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,7168,0.05099111133151584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,6144,0.047409776184293956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,5120,0.05836177534527249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,4096,0.028504000769721136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,4096,0.047741333643595375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,5120,0.037583112716674805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,4096,0.030985779232449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,3584,0.042344888051350914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,3584,0.025789333714379206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,3584,0.02815999918513828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,3072,0.03703822361098395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,3072,0.022964444425370958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,3072,0.02516266703605652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,2560,0.03154933452606201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,2560,0.020124443703227572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,2048,0.025987555583318073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,2560,0.022142221530278523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,2048,0.017321777012613084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,1536,0.02051911089155409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,2048,0.018732445107565988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,1536,0.015754666593339708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,1024,0.014527110589875115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,1024,0.012902222573757172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,768,0.010213333699438307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,768,0.011570666399266986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,768,0.011615110768212212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,512,0.008859555754396651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,512,0.008640000389681922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,512,0.010069333016872406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,256,0.005720888988839255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,256,0.006017777654859755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,256,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,128,0.004925333377387789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,128,0.003719999972316954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,128,0.006855999843941794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,64,0.0034088889757792154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,64,0.006852444261312485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,16384,32,0.0035822221802340615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,16384,32,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,65536,0.5506666501363119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,65536,0.292015102174547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,4096,0.09719289011425442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,16384,0.0733777748213874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,16384,0.14494400554233128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,16384,0.07608355416191949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,65536,0.3297742207845052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,12288,0.10875644286473592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,12288,0.06104711029264662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,65536,1536,0.043848888741599194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,10240,0.04836000005404154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,10240,0.09124000204934014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,8192,0.03992799917856852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,10240,0.05082577798101637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,8192,0.07418400049209595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,7168,0.0649991101688809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,8192,0.04189777705404493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,7168,0.037420445018344455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,6144,0.03169511093033685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,6144,0.056623107857174344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,6144,0.03324710991647508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,5120,0.027302222119437322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,5120,0.047247999244266085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,5120,0.02884888980123732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,4096,0.02288355595535702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,4096,0.039206223355399236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,4096,0.02405599918630388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,3584,0.0207013338804245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,3584,0.03456888927353753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,3584,0.02186577849917942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,3072,0.0185208883550432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,3072,0.030135999123255413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,3072,0.019647111495335896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,2560,0.016605334149466622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,2560,0.025808889004919264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,2560,0.017430222696728177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,16384,0.0935644441180759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,2048,0.014282666974597506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,2048,0.015097777048746744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,2048,0.02148977749877506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,1536,0.016787555482652452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,1536,0.012206222448084088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,1536,0.013222222526868185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,1024,0.012110222544935016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,1024,0.010685332947307162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,1024,0.011138666835096149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,768,0.009916444619496664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,768,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,512,0.007733333441946242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,512,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,512,0.008021333151393467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,256,0.004962666581074397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,256,0.0046737777690092725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,256,0.0064151109092765385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,128,0.004716444346639845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,128,0.003601777884695265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,128,0.006118222243256039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,64,0.0032471112079090546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,64,0.006160000132189856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,12288,32,0.0032808888289663526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,12288,32,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,5120,0.034141331911087036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,65536,0.2360008822547065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,1024,0.011702222128709158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,16384,1536,0.014608888162506951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,16384,0.11344267262352838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,16384,0.06758933597140841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,65536,0.4588444497850206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,12288,0.08630044592751397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,12288,0.04880977670351664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,16384,0.06704711251788668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,10240,0.07296178076002333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,10240,0.041787554820378624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,12288,0.05949422385957506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,10240,0.04503200120396084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,8192,0.059000889460245766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,65536,0.2621848848130968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,8192,0.037447998921076454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,7168,0.0520684454176161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,7168,0.03121244576242235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,6144,0.045808000697029956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,6144,0.02741777731312646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,7168,0.0373342235883077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,6144,0.029606223106384277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,5120,0.02380888826317257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,5120,0.03859377900759379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,4096,0.03178400132391188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,4096,0.01995022263791826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,5120,0.025760889053344727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,4096,0.021681777305073206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,3584,0.018533334136009216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,3584,0.028209778997633193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,3072,0.016602666841612924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,3584,0.01979377700222863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,3072,0.024875554773542616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,2560,0.021294222937689886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,3072,0.018028444714016385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,2560,0.014835556348164877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,2560,0.0159582214223014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,2048,0.017674666312005784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,2048,0.012882666455374824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,2048,0.014166222678290473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,1536,0.011234666738245221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,1536,0.013759999639458127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,1024,0.010240000155236986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,1536,0.012488889197508493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,1024,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,768,0.008546666966544257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,768,0.00847466703918245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,768,0.009776888622177972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,512,0.006807111203670502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,512,0.006342222293217977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,512,0.007227555745177799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,256,0.0037760001917680106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,256,0.006401777681377199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,256,0.005042666776312722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,128,0.0034373334298531213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,128,0.004825777891609403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,128,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,64,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,64,0.006211555666393704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,10240,32,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,10240,32,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,65536,0.3619724379645453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,16384,0.057985776000552706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,65536,0.2334702279832628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,16384,0.09009422196282281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,12288,0.05712622404098511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,16384,0.061517331335279674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,12288,0.06899733013576932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,12288,0.048469334840774536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,10240,0.03447022371821933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,7168,0.03575999869240655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,10240,0.05797866980234782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,10240,0.04165510998831855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,8192,0.028762665059831407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,8192,0.04727022184265984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,8192,0.03439822130733066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,7168,0.04184000028504265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,7168,0.031193776263131037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,6144,0.02299555473857456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,6144,0.03723911113209195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,6144,0.027641778190930683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,5120,0.020067556036843192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,5120,0.03120533294147915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,5120,0.02425866656833225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,4096,0.017156443662113614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,4096,0.025851555996470984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,4096,0.02022133270899455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,3584,0.015885333220163982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,3584,0.018498667412334018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,3584,0.02293599976433648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,3072,0.02022133270899455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,3072,0.014353776971499125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,12288,768,0.008955555657545725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,3072,0.01683911184469859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,2560,0.012841777669058906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,2560,0.01735911104414198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,2048,0.01442488862408532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,2560,0.014927110738224454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,2048,0.013350221845838757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,1536,0.009928888744778102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,1536,0.011429333024554782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,1536,0.011711111499203576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,1024,0.008353778057628209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,1024,0.008640888664457533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,1024,0.010031110710567897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,768,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,768,0.0075457774930530125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,768,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,512,0.005374222166008419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,512,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,512,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,256,0.004860444615284602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,256,0.0036640001667870414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,256,0.006340444501903322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,128,0.0033448889023727844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,128,0.004562666846646203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,64,0.0031519999934567344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,128,0.006050666587220297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,64,0.006043555421961679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,8192,32,0.0032497776879204642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,8192,32,0.006134222365087933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,65536,0.3384302192264133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,65536,0.17858310540517172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,16384,0.04759022262361315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,16384,0.08526755703820123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,65536,0.19621066252390543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,16384,0.04822133315934075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,12288,0.04079111085997687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,12288,0.061365332868364125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,12288,0.03844622108671401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,10240,0.031926221317715116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,10240,0.0517502228418986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,10240,0.03313866588804457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,8192,0.041983111037148364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,8192,0.02681511143843333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,8192,0.03479199939303928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,7168,0.03741066654523214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,8192,0.027544000082545813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,7168,0.024007999234729346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,6144,0.033195555210113525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,7168,0.025111999776628282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,6144,0.021583999196688335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,6144,0.022678222921159532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,5120,0.028106666273540918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,5120,0.018797333041826885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,4096,0.02330133318901062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,5120,0.020048000746303134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,4096,0.016172443826993305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,3584,0.020850666695170935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,4096,0.01684533390733931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,3584,0.01475288967291514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,3584,0.015591111448076038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,3072,0.018481777773963082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,3072,0.013391999734772576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,10240,1024,0.00944088896115621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,3072,0.014370666609870063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,2560,0.015836444165971544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,2560,0.012151111331250934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,2560,0.013146667016877068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,2048,0.013028444515334235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,2048,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,1536,0.009511111511124505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,1536,0.010607110957304636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,1536,0.01038755559258991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,1024,0.00813688917292489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,1024,0.008054222497675154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,1024,0.008858666651778752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,768,0.0064444442590077715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,768,0.0069439998931354946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,768,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,512,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,512,0.005438222239414851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,512,0.006118222243256039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,256,0.0047182221379545005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,256,0.0034871109657817413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,256,0.005717333406209946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,128,0.0032346666687064697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,128,0.004521777646409141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,128,0.005524444497293896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,64,0.003024000053604444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,64,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,7168,32,0.0030453333424197305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,7168,32,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,65536,0.27301867802937824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,65536,0.14639466338687473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,16384,0.04116800096299913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,65536,0.18464977211422393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,16384,0.07882133457395765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,65536,0.17945155832502577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,16384,0.05010399884647793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,12288,0.03261422117551168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,12288,0.05745689074198405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,12288,0.034487111700905695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,10240,0.04672266708479988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,10240,0.027887112564510767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,12288,0.04053511222203573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,8192,0.03899022274547153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,10240,0.030157334274715845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,8192,0.0233262214395735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,8192,0.02568000058333079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,7168,0.020994666549894545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,7168,0.03432088759210374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,7168,0.023249778482649062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,6144,0.01885422236389584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,6144,0.030231111579471167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,6144,0.02095111045572493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,5120,0.016696888539526198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,5120,0.025664000047577754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,5120,0.018545778261290658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,4096,0.02160177793767717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,4096,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,4096,0.015822221835454304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,3584,0.019276445110638935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,3584,0.014551111393504672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,3072,0.012327111429638334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,3072,0.01346577786737018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,3072,0.018529777725537617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,3584,0.014733332726690503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,2560,0.014679110712475248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,2560,0.011177777416176267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,2560,0.012362666428089142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,2048,0.01108266661564509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,2048,0.012475555141766867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,1536,0.00888888869020674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,1536,0.010112889111042023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,1536,0.009886222581068674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,1024,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,7168,0.025819554924964905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,1024,0.007600888609886169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,1024,0.007298666569921706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,768,0.006443555570311016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,768,0.005990222096443176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,512,0.004553777890072929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,768,0.0064106666379504735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,512,0.006045333213276333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,512,0.005120000077618493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,256,0.0047315553658538395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,256,0.0034551111360390982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,256,0.005655999812814925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,128,0.005460444423887465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,128,0.0032088889016045462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,128,0.004536889079544279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,64,0.005466666486528184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,64,0.0029964444951878656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,6144,32,0.003026666740576426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,6144,32,0.005500444521506627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,16384,0.0637635588645935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,16384,0.03756444321738349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,65536,0.22717777887980142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,12288,0.02962755494647556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,16384,0.03712000118361579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,12288,0.05201155609554715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,10240,0.02545333405335744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,12288,0.030143999391131934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,10240,0.03826844361093309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,65536,0.12781866391499838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,10240,0.02629511058330536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,8192,0.03132889005872939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,8192,0.02148444453875224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,7168,0.02797422144148085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,8192,0.02421955598725213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,7168,0.019606222709019978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,7168,0.020224889119466145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,6144,0.024941333466105994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,6144,0.01827822294500139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,5120,0.021327111456129286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,5120,0.015672889020707872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,5120,0.01646311084429423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,4096,0.013618666264745923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,4096,0.017654221918847825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,4096,0.013985777894655863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,3584,0.012482666307025485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,3584,0.015779554843902588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,3584,0.013178666432698568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,3072,0.011459555890825061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,3072,0.014001778430408902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,8192,2048,0.01126488877667321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,3072,0.012223111258612739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,2560,0.01051288925939136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,2560,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,2560,0.01128266668981976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,2048,0.010279999838935005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,2048,0.009319999979601966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,1536,0.008657777474986183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,2048,0.010314666562610203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,1536,0.008459555606047312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,1536,0.009328888522254096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,1024,0.006862222320503659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,1024,0.006405333264006509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,1024,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,768,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,768,0.006171555568774541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,512,0.0038399998512532976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,512,0.004979555391603046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,512,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,256,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,256,0.0033573332346147964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,256,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,128,0.0031235555393828284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,128,0.004401777767472797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,128,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,64,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,64,0.005483555710977978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,5120,32,0.002944888960984018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,5120,32,0.00544177782204416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,65536,0.18204800287882486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,65536,0.10821511348088582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,16384,0.030470222234725952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,16384,0.050711999336878456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,65536,0.11407111750708686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,16384,0.03201955556869507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,12288,0.024176888995700415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,12288,0.04046577877468533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,12288,0.026377777258555096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,10240,0.0210506667693456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,10240,0.03429333368937174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,10240,0.02295822236273024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,8192,0.02877955635388692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,8192,0.019318222999572754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,7168,0.016294222739007738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,7168,0.02614044480853611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,7168,0.017837334010336135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,6144,0.015017777681350708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,6144,0.020943111843532983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,6144,0.016249777542220224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,5120,0.013389333254761167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,5120,0.01792888840039571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,5120,0.014733332726690503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,4096,0.0147733340660731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,4096,0.011809777882364062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,4096,0.012642666697502136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,3584,0.011166222393512726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,3584,0.013157332936922709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,3584,0.011898666620254517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,3072,0.010291555689440833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,3072,0.011887999872366587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,3072,0.01111644423670239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,2560,0.010327999790509542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,2560,0.009494221872753566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,2560,0.010418666733635796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,2048,0.00851111114025116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,2048,0.008886222210195329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,2048,0.009574222067991892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,1536,0.007375111182530721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,1536,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,1536,0.008182221816645728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,1024,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,1024,0.005717333406209946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,1024,0.006225777582989798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,768,0.0047022220161226065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,768,0.005507555686765247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,768,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,512,0.004968889057636261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,512,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,512,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,256,0.003228444399105178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,256,0.004558222161398994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,256,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,128,0.004424889054563311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,128,0.0029680000411139596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,128,0.005323555734422471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,64,0.0028373334142896864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,64,0.00534400012758043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,4096,32,0.00299822228650252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,4096,32,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,7168,2048,0.010781333678298526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,65536,0.16963999801211885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,65536,0.09964978032641941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,16384,0.029673778348498877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,16384,0.04568799999025133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,16384,0.02971733278698391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,65536,0.10235022173987494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,12288,0.02236088944805993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,12288,0.03618933426009284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,12288,0.024474667178259954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,10240,0.01990133358372582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,10240,0.031174222628275555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,10240,0.02158755560715993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,8192,0.01680888897842831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,8192,0.026171555121739704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,8192,0.01795111099878947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,7168,0.015488000379668342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,7168,0.0236799998415841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,7168,0.01683911184469859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,6144,0.014160888062583076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,6144,0.018743111027611625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,6144,0.01536177761024899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,5120,0.012677333421177335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,5120,0.015944000747468736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,5120,0.013997332917319404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,4096,0.013159111142158508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,4096,0.01125511113140318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,4096,0.011991110940774282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,3584,0.01054044481780794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,3584,0.011896889242861005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,3584,0.011376000112957425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,3072,0.009862222605281407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,3072,0.010796444283591377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,3072,0.010775110787815519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,2560,0.009090666969617208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,2560,0.009566222627957663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,2560,0.010035555395815108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,2048,0.008259555531872643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,2048,0.008291555775536431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,2048,0.009213333328564962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,1536,0.006441777778996362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,1536,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,1536,0.007093333535724216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,1024,0.005524444497293896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,1024,0.005800888770156437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,1024,0.00620888877246115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,768,0.004366222355100843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,768,0.005435555759403441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,768,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,512,0.0035777779089079965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,512,0.004863111095296012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,512,0.005751111027267244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,256,0.0031679999083280563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,256,0.005493333356248007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,256,0.004588444613748127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3584,128,0.004301333179076512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,64,0.0027599999060233435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,128,0.005370666583379109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,128,0.0029226665695508323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,64,0.005349333501524395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3584,32,0.0028488888508743713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3584,32,0.005411555369695027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,65536,0.14052266544765896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,16384,0.028375112348132666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,65536,0.08855644199583267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,16384,0.042154666450288564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,6144,2048,0.009908444351620143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,16384,0.02760266595416599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,12288,0.021431111627154883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,12288,0.03327288892534044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,12288,0.022628444764349196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,10240,0.028392000330819026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,10240,0.019095111224386428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,10240,0.019941333267423842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,8192,0.024363555841975745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,8192,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,7168,0.014903111590279473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,7168,0.022015111313925848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,7168,0.01571111049917009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,6144,0.013703111145231458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,6144,0.01953155630164676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,6144,0.014575999644067554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,5120,0.012263110942310758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,5120,0.01722666621208191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,5120,0.013371555341614617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,4096,0.010958222051461538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,4096,0.015037332971890768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,4096,0.011526222030321756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,3584,0.010978666444619497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,3584,0.01015466699997584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,3584,0.011081777513027191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,3072,0.009925333162148794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,3072,0.0096124443742964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,3072,0.010405333505736457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,2560,0.008767999708652496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,2560,0.009004444711738164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,2560,0.009708444277445475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,2048,0.00777688870827357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,65536,0.14200354946984187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,2048,0.007894222107198503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,2048,0.008492444124486711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,1536,0.006867555694447623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,6144,0.017633777525689866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,1536,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,1024,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,1024,0.005510222166776657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,1024,0.006179555422729916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,768,0.005347555710209741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,768,0.003918222255176968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,768,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,512,0.004844444493452708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,512,0.00572533326016532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,512,0.003579555700222651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,256,0.003212444484233856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,256,0.005489777773618698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,256,0.004463110946946674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,128,0.0029013332807355454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,128,0.004230222354332606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,128,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,64,0.0028044444819291434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,64,0.005336888962321811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,3072,32,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,3072,32,0.005388444496525659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,65536,0.11949333879682754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,65536,0.0805564456515842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,5120,768,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,16384,0.03450399968359206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,16384,0.023333332604832117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,65536,0.08091200060314603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,16384,0.02569599946339925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,12288,0.01880088945229848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,12288,0.02769600020514594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,12288,0.02106222179200914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,10240,0.017567111386193168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,10240,0.024030221833123103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,10240,0.018767111831241183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,8192,0.014854222536087036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,8192,0.020317332612143624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,8192,0.015948444604873657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,7168,0.0184515549076928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,7168,0.013760888742076026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,7168,0.015032000011867948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,6144,0.012681778106424542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,6144,0.016525333126386006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,6144,0.013842666314707862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,5120,0.011549333731333414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,5120,0.014583999911944071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,5120,0.012884444660610624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,4096,0.012637333737479316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,4096,0.010295111272070143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,4096,0.011135111252466837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,3584,0.011893333660231696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,3584,0.010687111152542962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,3072,0.009000889129108852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,3072,0.00908622228437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,3072,0.010017777482668558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,2560,0.008203555312421586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,2560,0.008254222571849823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,2560,0.009121778110663096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,2048,0.00721155520942476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,2048,0.006811555474996567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,2048,0.0075031113293435835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,1536,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,1536,0.0063884444534778595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,1536,0.006520888871616787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,1024,0.0047306666771570844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,1024,0.005402666827042897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,1024,0.006072000082996156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,768,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,768,0.0038817777401871155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,768,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,512,0.004754666652944353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,512,0.003571555432346132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,512,0.005664000080691443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,256,0.004412444515360726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,256,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,256,0.005490666462315454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,128,0.0029226665695508323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,128,0.004222222086456087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,64,0.002749333365095986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,128,0.0053226666318045715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,64,0.0053102220926019884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2560,32,0.00282222218811512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2560,32,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,65536,0.09746844238705105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,65536,0.07872888776991102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,16384,0.022680888573328655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,65536,0.07514666848712497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,16384,0.028560890091790095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,16384,0.02404800057411194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,12288,0.018041777941915724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,12288,0.023205333285861548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,4096,8192,0.017833777599864535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,12288,0.019680889116393197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,10240,0.02049333353837331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,10240,0.016082665986484952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,10240,0.017646221650971305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,8192,0.01695999999841054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,8192,0.015172445111804538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,8192,0.013844444519943662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,7168,0.016506666938463848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,7168,0.012952889005343119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,7168,0.014204444156752693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,6144,0.011914666328165265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,6144,0.014038221703635322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,6144,0.013180444637934366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,5120,0.01258488910065757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,5120,0.010727110836240979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,5120,0.012143111063374413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,4096,0.011157333023018308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,4096,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,4096,0.010518222219414182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,3584,0.01050222251150343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,3584,0.009062222308582729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,3584,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,3072,0.008480889101823172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,3072,0.008619555168681675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,3072,0.009279111193286048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,2560,0.007567110988828871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,2560,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,2560,0.007878222399287755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,2048,0.0064791109826829694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,2048,0.006872889068391588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,2048,0.006712889091836081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,1536,0.005514666852023866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,1536,0.0063768890168931745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,1024,0.005418666534953647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,1024,0.004101333518822988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,1024,0.005994666781690385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,768,0.005392000079154968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,768,0.0038195554580953387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,768,0.005767999837795894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,512,0.0034044444974925783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,512,0.0047893333766195504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,512,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,256,0.004507555729813046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,256,0.0030577776746617425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,256,0.005405333307054307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,128,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,128,0.00435288871328036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,128,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,64,0.0026933333526055017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,2048,32,0.0027066665805048416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,64,0.0052328887912962176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,2048,32,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,65536,0.07612888680564033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,65536,0.06828622023264568
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,16384,0.020940444535679285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,16384,0.023375999596383836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,65536,0.0703279972076416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,16384,0.02241955531968011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,12288,0.01699733402993944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,12288,0.018836443622907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,12288,0.0184879998366038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,10240,0.015293333265516492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,10240,0.016806221670574613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,10240,0.016807110773192514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,8192,0.013637333280510373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,8192,0.014239999983045789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,8192,0.014383110735151501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,7168,0.012487110992272695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,7168,0.013203555511103736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,7168,0.013592888911565145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,6144,0.011582222249772815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,6144,0.012134222520722283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,6144,0.012675555215941535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,5120,0.011095999843544431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,5120,0.011716444459226398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,4096,0.00905600024594201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,4096,0.00922400007645289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,4096,0.010046222143703038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,3584,0.00870666652917862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,3584,0.00851288851764467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,3584,0.009304888546466827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,3072,0.007527111305130853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,3072,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,3072,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,2560,0.007207999626795451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,2560,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,2560,0.00720266666677263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,2048,0.006001777946949005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,2048,0.006337777607970768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,2048,0.006741333338949416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,1536,0.005783111270931032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,1536,0.004929777648713854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,1536,0.006445333361625671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,1024,0.005203555441564984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,1024,0.0041315555572509766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,1024,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,768,0.005119110975000593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,768,0.0037368887828456033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,768,0.0058666666348775225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,512,0.0046844445168972015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,512,0.0034924443397257063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,512,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,256,0.004395555704832077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,256,0.0030417777597904205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,256,0.005424889011515512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,128,0.004176889028814104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,128,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,128,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,64,0.002671111168132888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,64,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1536,32,0.0027457777824666765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1536,32,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,65536,0.06726844443215264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,65536,0.05420355664359199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,65536,0.06726133161120944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,16384,0.020388444264729817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,16384,0.017625777257813346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,16384,0.02142933342191908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,12288,0.01643111142847273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,12288,0.014410666293568082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,12288,0.017667555146747164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,10240,0.01460000044769711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,10240,0.013002666334311167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,10240,0.016029333074887592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,8192,0.012701333396964602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,8192,0.01142133358452055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,8192,0.013904000322024027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,7168,0.01183644433816274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,7168,0.010768888725174798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,7168,0.013040888640615674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,65536,0.09099288781483968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,6144,0.01035733355416192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,6144,0.00979911122057173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,6144,0.011719110939237805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,5120,0.009296889106432596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,5120,0.010012444522645738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,5120,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,4096,0.008001777860853408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,4096,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,4096,0.009278222090668148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,3584,0.007383999725182851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,3584,0.00872444444232517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,3584,0.00813600007030699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,3072,0.008188444707128737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,3072,0.007557333343558841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,2560,0.006488889041874144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,2560,0.006539555473460092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,2560,0.007176888485749562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,2048,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,2048,0.005434666656785541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,8192,0.016185777054892648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,2048,0.006781333436568578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,1536,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,1536,0.004807999978462855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,1536,0.006413333531883027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,1024,0.0041031113101376426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,1024,0.005091555416584015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,1024,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,768,0.0037084445357322693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,768,0.0058133333093590206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,512,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,512,0.0034417777011791864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,512,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,256,0.004395555704832077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,256,0.003063111048605707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,256,0.0053635554181204895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,128,0.0028408887899584244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,128,0.0041893331540955436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,128,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,64,0.0027066665805048416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,1024,32,0.0026968889352348116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,64,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,1024,32,0.005212444398138259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,65536,0.04400177796681722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,3072,1536,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,65536,0.06749066379335192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,16384,0.020488889680968392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,16384,0.01479644411140018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,65536,0.06632266441980998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,16384,0.021536888347731695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,12288,0.01643111142847273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,12288,0.012451555165979596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,12288,0.01777688827779558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,10240,0.014535110857751636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,10240,0.011605333122942181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,10240,0.017651556266678702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,8192,0.010318222145239512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,8192,0.012679999901188744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,8192,0.016720889343155753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,7168,0.010662222074137794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,7168,0.011542222566074796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,7168,0.01555377741654714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,6144,0.010125333236323463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,6144,0.00998222248421775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,6144,0.013859555953078799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,5120,0.0092613332801395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,5120,0.00871288859181934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,5120,0.011990221838156382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,4096,0.008387555678685507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,4096,0.00777422222826216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,4096,0.010432889064153036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,3584,0.007391110890441471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,3584,0.007599111232492659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,3584,0.009927111367384592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,3072,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,3072,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,3072,0.009762666291660732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,2560,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,2560,0.006401777681377199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,2560,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,2048,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,2048,0.0053591111467944244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,2048,0.006772444479995304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,1536,0.004655999855862724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,1536,0.005522666705979242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,1536,0.0063440000845326324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,1024,0.005096000101831225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,1024,0.003997333347797394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,1024,0.006020444548792309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,768,0.0047768888374169665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,768,0.0036959999965296853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,768,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,512,0.004620444443490771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,512,0.0033448889023727844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,512,0.00563466673096021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,256,0.0042257776690853965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,256,0.0030293334275484085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,256,0.005373333477311664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,768,128,0.004159111115667555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,128,0.0027724444452259275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,128,0.005326222214433882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,64,0.0026568888376156488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,768,32,0.002647111192345619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,64,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,768,32,0.0052915554907586836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,65536,0.032771554258134626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,65536,0.0675555533832974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,16384,0.020204444726308186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,65536,0.06589333216349284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,16384,0.011874666644467248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,16384,0.021184888150956895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,12288,0.016303999556435477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,12288,0.011398221883508893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,12288,0.017521777086787753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,10240,0.010616000327799054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,10240,0.014133334159851074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,10240,0.015241778559154935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,8192,0.011734222372372946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,8192,0.00960977789428499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,8192,0.012647111382749347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,7168,0.010677333507272931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2560,3584,0.009764444496896531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,7168,0.009122666385438707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,7168,0.011045333411958484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,6144,0.009640889035330879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,6144,0.00833333366447025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,6144,0.010207999911573198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,5120,0.0074444446298811175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,5120,0.009444444543785518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,5120,0.008692444198661381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,4096,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,4096,0.008393777741326226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,4096,0.007738666401969061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,3584,0.007400888535711501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,3584,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,3584,0.0069217777086628815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,3072,0.006510222123728857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,3072,0.007603555917739868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,3072,0.0068400001360310455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,2560,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,2560,0.0063493334584765965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,2560,0.007221333682537079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,2048,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,2048,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,2048,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,1536,0.005438222239414851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,1536,0.00461777796347936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,1536,0.006396444307433234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,1024,0.003944000022278892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,1024,0.005035555611054103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,1024,0.005998222364319696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,768,0.003589333345492681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,768,0.0058062221441004015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,768,0.004856889032655292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,512,0.0033333332588275275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,512,0.004688888788223267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,512,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,256,0.0029360000044107437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,256,0.005375111268626318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,256,0.004271111140648524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,128,0.002728888971938027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,128,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,64,0.002583111118939188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,64,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,512,32,0.00258933338854048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,512,32,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,65536,0.022132444712850783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,65536,0.06587821907467313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,16384,0.019280888968043856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,16384,0.00997155573632982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,16384,0.02016533414522807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,12288,0.015238222148683337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,12288,0.008897778060701158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,12288,0.016081778539551627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,10240,0.013306666579511432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,10240,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,10240,0.014295111099878946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,2048,1536,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,8192,0.007183111376232571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,8192,0.011904000408119626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,7168,0.010605333579911126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,7168,0.006680000159475539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,7168,0.011126221881972419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,6144,0.009693333672152625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,6144,0.0064080001579390625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,6144,0.010282666318946415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,5120,0.008763555851247575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,5120,0.00685511115524504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,5120,0.009507555928495195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,4096,0.007822222179836696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,4096,0.0064088888466358185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,4096,0.008416000339719985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1536,5120,0.01052266690466139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,3584,0.006852444261312485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,3584,0.007997333175606197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,3072,0.006785777707894643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,3072,0.006537777682145436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,3072,0.0075786668393347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,2560,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,2560,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,2560,0.00721777809990777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,2048,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,2048,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,2048,0.006720888945791457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,1536,0.005432888865470886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,1536,0.004552000098758274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,1536,0.006395555618736479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,1024,0.003949333396222857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,1024,0.005988444305128521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,768,0.00481599983241823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,768,0.0035982223020659555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,768,0.005792888916201062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,512,0.004480888860093223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,512,0.003244444520937072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,512,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,256,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,256,0.00295733329322603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,256,0.005375111268626318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,128,0.00407911092042923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,128,0.0027511111564106415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,128,0.0052595556610160405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,64,0.0026488889836602737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,64,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,256,32,0.0026551110463009942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,256,32,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,65536,0.06713689035839505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,65536,0.015929778416951496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,65536,0.0662826697031657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,16384,0.019023999571800232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,16384,0.007099555598364935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,16384,0.019294222195943195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,12288,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,12288,0.006720888945791457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,12288,0.01537333263291253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,10240,0.01332800007528729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,10240,0.00674044465025266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,10240,0.013638222383128272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,8192,0.011343110766675739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,8192,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,8192,0.011744000017642975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,7168,0.010464000205198923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,7168,0.006518222391605377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,7168,0.011165333290894827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,6144,0.009573332965373993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,6144,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,6144,0.01016711112525728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,5120,0.00866844422287411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,5120,0.006683555742104848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,5120,0.009401777552233802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,4096,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,4096,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,4096,0.008316444853941599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,3584,0.007188444336255391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,3584,0.00676711110605134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,3584,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,3072,0.006990222467316522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,3072,0.006767999794748094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,3072,0.006457777900828256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,3072,0.00757244477669398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,2560,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,2560,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,2560,0.007195555501514011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,2048,0.005766222046481238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,2048,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,2048,0.006728000111050076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,1536,0.005324444423119227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,1536,0.004605333424276776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,1536,0.006365333166387346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,1024,0.003938666648334927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,768,0.0046986668474144405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,768,0.00361333332127995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,768,0.005771555420425203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,512,0.004467555632193883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,512,0.003289777785539627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,512,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,256,0.0041928887367248535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,1024,768,0.004905777672926585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,256,0.002973333415057924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,256,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,128,0.0027715555495686005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,128,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,64,0.0026142222599850762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,64,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,128,32,0.0026373333401150177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,128,32,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,65536,0.014188443620999655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,65536,0.06657155354817708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,16384,0.006522666662931442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,16384,0.01893422173129188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,12288,0.0064088888466358185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,12288,0.01534933348496755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,10240,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,10240,0.013728888498412238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,8192,0.006490666833188798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,8192,0.011919111013412476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,7168,0.0064328888224230866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,7168,0.01111822244193819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,6144,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,6144,0.010286221901575724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,5120,0.006684444430801604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,5120,0.009525333013799455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,4096,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,4096,0.008476444416575963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,3584,0.006781333436568578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,3584,0.008058666355080074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,3072,0.0064399999876817065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,3072,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,2560,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,2560,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,2048,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,2048,0.006809777683681912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,1536,0.004661333229806689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,1536,0.0064186664919058485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,512,128,0.004084444294373195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,1024,0.003972444269392225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,768,0.0035955554081334006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,768,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,512,0.0032417778339650896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,512,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,256,0.002893333426780171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,256,0.005330666485759947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,128,0.0027475555737813315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,128,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,64,0.0026071110947264563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,64,0.0052168890833854675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,64,32,0.0026346666531430352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,64,32,0.005179555465777715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,65536,0.012890666723251343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,16384,0.006500444478458828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,65536,0.06589421961042616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,16384,0.018756444255510967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,12288,0.015303999185562134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,12288,0.0063271112740039825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,10240,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,10240,0.013540444274743399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,8192,0.0064604443808396654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,8192,0.011721777419249216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,7168,0.006380444599522485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,7168,0.010963555839326648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,6144,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,6144,0.010155555274751451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,5120,0.006660444455014334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,5120,0.009382222261693742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,4096,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,4096,0.008303110798199972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,3584,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,3584,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,3072,0.006321777900060018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,3072,0.007575111256705389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,2560,0.0058782220714622075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,2560,0.007166222565703922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,2048,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,2048,0.006748444504208035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,1536,0.004524444540341695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,1024,0.00387999994887246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,1536,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,768,0.003528888854715559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,1024,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,512,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,768,0.005753777921199799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,512,0.005497777627574072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,256,0.0028773333049482773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,128,0.0026968889352348116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,256,0.0053013331360287136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,64,0.0025457777082920074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,128,0.0051857779423395796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,64,0.005155555489990446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,48,32,32,0.0025555555605226094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,48,32,32,0.005162666655249066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,65536,0.06675733460320367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,16384,0.7181270917256674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,8192,0.01146666705608368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,16384,0.37839023272196454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,3584,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,12288,0.5498879750569662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,16384,0.3925786548190647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,12288,0.3133457766638862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,10240,0.48869779374864364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,10240,0.24078933397928873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,12288,0.28034665849473744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,8192,0.366592009862264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,10240,0.270592901441786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,256,1024,0.004914666629499859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,8192,0.2004044453303019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,7168,0.3168275621202257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,7168,0.18481511539883086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,6144,0.27376890182495117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,7168,0.174563553598192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,6144,0.1387946605682373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,5120,0.11652889516618516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,5120,0.22827021280924478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,6144,0.15404621760050455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,5120,0.12823822763231066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,4096,0.09665511051813762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,4096,0.18547466066148546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,3584,0.16172266006469727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,3584,0.08726755777994792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,4096,0.1162853373421563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,3584,0.09042933252122666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,3072,0.07698133256700304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,3072,0.13971288998921713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,2560,0.0662248863114251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,2560,0.11729155646430121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,3072,0.07658311393525866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,2560,0.06545777453316583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,2048,0.059248891141679555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,2048,0.09507199790742661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,1536,0.07236533032523261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,1536,0.041032890478769936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,2048,0.05428444345792135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,1536,0.043030222256978355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,1024,0.03188800149493747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,1024,0.05225511060820686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,768,0.02402755618095398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,1024,0.03151911165979173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,768,0.039176887936062284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,512,0.02773422168360816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,512,0.019009777241282992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,768,0.02869422237078349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,256,0.013937777943081327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,512,0.02102488941616482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,256,0.015604444675975375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,256,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,128,0.009303111169073317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,64,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,128,0.014175110393100314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,65536,32,0.005209777918126848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,64,0.011074666347768573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,65536,32,0.011235555840863122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,1024,0.004922666483455234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,16384,0.17587822013431123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,16384,0.09548622369766235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,48,128,128,0.003987555702527364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,12288,0.07268089056015015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,12288,0.13415822717878553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,65536,0.7040195465087891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,16384,0.11249066723717584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,10240,0.13121599621242946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,10240,0.06521066692140368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,12288,0.08028711213005914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,10240,0.07099111212624444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,8192,0.09056266811158921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,8192,0.05053866571850247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,7168,0.07975466383828057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,7168,0.0451217790444692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,8192,0.05625333388646444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,6144,0.06969955232408312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,65536,0.37691733572218156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,7168,0.04989688926272922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,65536,0.43002843856811523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,6144,0.043980442815356784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,5120,0.034338666333092585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,4096,0.0473653342988756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,5120,0.061787552303738065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,5120,0.0380604432688819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,4096,0.029397332006030615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,4096,0.031352887550989784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,3584,0.02576711111598545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,3584,0.044473777214686074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,3584,0.028394665982988145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,3072,0.03672177924050225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,3072,0.02383733292420705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,2560,0.031162665949927434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,3072,0.025223111112912495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,2560,0.022154667311244543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,2560,0.020767110917303298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,2048,0.025504888759719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,2048,0.017866666118303936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,2048,0.018938667244381376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,1536,0.021036444438828364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,1536,0.01571111049917009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,1024,0.014259555273585849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,1024,0.011685333318180509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,768,0.009974222216341231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,768,0.01128000020980835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,1024,0.013457777599493662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,512,0.008520888785521189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,512,0.00906044410334693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,768,0.011972444752852121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,512,0.010086221827401055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,256,0.006135111053784688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,256,0.007054222126801808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,128,0.003583999971548716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,256,0.005439110928111606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,128,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,128,0.0047315553658538395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,64,0.0033288889875014624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,64,0.006704888823959563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,16384,32,0.0033679999825027254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,16384,32,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,65536,0.5416053136189779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,16384,0.08484177456961738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,8192,0.18151022328270805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,65536,0.2852213382720947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,65536,0.3109235498640272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,16384,0.1403475602467855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,16384,0.077802664703793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,12288,0.11698577139112686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,12288,0.059841778543260365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,10240,0.0914577775531345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,10240,0.04784444305631849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,10240,0.052629331747690834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,8192,0.07466933462354872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,8192,0.03954666521814134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,8192,0.04231377773814731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,7168,0.06524178054597643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,7168,0.03533066643608941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,6144,0.056518223550584584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,7168,0.03815999958250258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,6144,0.031211555004119873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,6144,0.03336800138155619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,5120,0.027015111512607996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,5120,0.04708977871470981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,4096,0.03903288973702325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,5120,0.029166221618652344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,3584,0.03433155682351854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,3584,0.020498666498396132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,4096,0.02439022229777442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,3072,0.029855112234751385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,3072,0.018364444375038147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,3584,0.022061333060264587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,2560,0.025633777181307476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,3072,0.020059555768966675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,2560,0.016352888610627916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,2560,0.017424000634087455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,2048,0.021116443806224402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,2048,0.014223999447292753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,1536,0.016700444949997794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,2048,0.015203555425008139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,1536,0.01204355557759603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,1024,0.011847111086050669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,1536,0.013188444077968597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,1024,0.009926222264766693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,1024,0.011101333631409539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,768,0.009653333160612319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,768,0.008832889298597971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,768,0.01015377789735794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,512,0.007365333537260692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,512,0.007423111134105259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,512,0.008065777520338694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,256,0.004361777669853634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,256,0.004828444371620814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,256,0.00638755535085996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,128,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,128,0.00443822228246265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,128,0.006044444524579578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,64,0.0031617778456873367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,12288,32,0.0031866667171319327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,64,0.006125333408514659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,12288,32,0.0061919999619325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,65536,0.45517778396606445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,65536,128,0.011222222612963783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,16384,0.11404800415039062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,16384,0.07167288992140028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,65536,0.23925688531663683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,65536,0.2659822305043538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,16384,0.06880799929300944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,12288,0.04924799998601278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,12288,0.08700088659922282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,12288,0.059952000776926674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,10240,0.07311022281646729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,10240,0.04343200061056349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,10240,0.04511022236612108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,8192,0.05927822324964735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,8192,0.0348693331082662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,7168,0.0521653327676985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,7168,0.03128977616628011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,8192,0.03940622342957391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,6144,0.027768888407283362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,7168,0.03343288766013251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,6144,0.04589866598447164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,6144,0.03189511100451151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,5120,0.03867910967932807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,5120,0.024336000283559162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,5120,0.025853332546022203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,4096,0.03156266609827677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,4096,0.02182666626241472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,3584,0.018593778212865193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,3584,0.028037332826190527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,3584,0.019928889142142404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,3072,0.016929777132140267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,3072,0.024656888511445787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,3072,0.017980444762441847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,2560,0.014967999524540372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,2560,0.021094222863515217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,2560,0.01591466698381636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,2048,0.012967111335860359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,2048,0.017384888397322763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,2048,0.014070222775141398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,1536,0.013644444445768992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,1536,0.011236444115638733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,1536,0.01240711079703437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,1024,0.009278222090668148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,1024,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,1024,0.010674667027261523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,768,0.008436444732877944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,768,0.008348444269763099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,768,0.009756444229020013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,512,0.006223111102978389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,512,0.006687110910813014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,512,0.007045333584149678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,256,0.0036844445599450003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,256,0.004913777940803104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,256,0.006343110981914733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,6144,0.039629333549075656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,128,0.0032497776879204642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,128,0.004536889079544279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,64,0.002982222164670626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,128,0.006130666782458623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,64,0.006161777923504512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,10240,32,0.003036444385846456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,10240,32,0.006200888918505774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,16384,1536,0.014553777045673795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,16384,0.050957332054773964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,65536,0.35872356096903485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,16384,0.0892346633805169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,16384,0.06715644730461968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,65536,0.18887021806504992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,65536,0.22963645723130968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,12288,0.06856977939605713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,10240,0.057497779528299965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,12288,0.05353422297371758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,10240,0.041688889265060425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,8192,0.028705779049131606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,8192,0.0468355549706353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,8192,0.03379999929004245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,7168,0.02565866708755493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,7168,0.04153066542413499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,7168,0.030870222383075293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,6144,0.0366213321685791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,6144,0.023059555225902136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,6144,0.02740977704524994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,5120,0.030968000491460163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,5120,0.020035554965337116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,5120,0.023932443724738225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,4096,0.025546666648652818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,4096,0.017147555947303772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,4096,0.019951999187469482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,3584,0.02272266646226247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,3584,0.01588888963063558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,3584,0.018395556343926322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,3072,0.020022221737437777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,3072,0.016708445217874315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,2560,0.017229333519935608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,2560,0.012880889077981314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,2560,0.014936000108718872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,2048,0.011344888971911537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,2048,0.014105778601434497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,1536,0.0098213329911232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,2048,0.01333866682317522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,1536,0.011339555184046427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,1536,0.011575111084514193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,1024,0.008367111285527548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,1024,0.008546666966544257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,1024,0.010051555931568146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,768,0.007233777807818518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,768,0.007207999626795451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,768,0.008040889269775814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,512,0.0057911111248864075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,512,0.005179555465777715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,512,0.00676711110605134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,256,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,256,0.0046968890560997855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,256,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,128,0.0031999999450312722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,128,0.004363555461168289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,64,0.003012444410059187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,128,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,64,0.006055999961164262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,8192,32,0.003063999944263034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,8192,32,0.006070222291681502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,12288,0.056658665339152016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,65536,0.3215564356909858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,12288,4096,0.02241866621706221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,16384,0.052932444545957774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,16384,0.07966933647791545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,65536,0.17111289501190186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,65536,0.17819378111097547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,16384,0.04923289020856222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,12288,0.06527288754781087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,12288,0.042041778564453125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,10240,0.05169511172506544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,10240,0.03277066681120131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,12288,0.038605332374572754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,8192,0.04182489050759209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,10240,0.03347466720475091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,8192,0.026760000321600173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,8192,0.027808000644048054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,7168,0.024012444747818842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,7168,0.03712977634535895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,6144,0.0326275560590956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,7168,0.025224000215530396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,6144,0.021584888299306233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,6144,0.022629333866967097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,5120,0.027868443065219458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,5120,0.02026844521363576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,4096,0.016103999482260812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,4096,0.023056889573733013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,4096,0.016843555702103507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,3584,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,3584,0.020582222276263766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,3584,0.015657777587572735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,3072,0.013432000246312885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,3072,0.018287110659811232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,3072,0.014416888356208801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,2560,0.0122079998254776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,2560,0.01573777695496877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,2560,0.013018666870064206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,2048,0.012963555753231049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,2048,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,2048,0.01165511127975252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,1536,0.009462222456932068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,1536,0.0105591110057301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,1536,0.01036177741156684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,1024,0.008071111308203803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,1024,0.008034666379292807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,1024,0.00889866633547677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,768,0.006813333266311222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,768,0.006232000059551663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,768,0.007119110888904995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,512,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,512,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,512,0.006076444354322221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,256,0.00462755560874939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,256,0.003363555504216088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,256,0.005737777799367905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,128,0.003104000041882197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,128,0.004330666528807746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,64,0.0030782222747802734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,128,0.005471111171775394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,7168,32,0.0029493332323100832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,64,0.005489777773618698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,7168,32,0.005493333356248007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,65536,0.26878489388359916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,65536,0.1460097763273451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,10240,4096,0.02014311154683431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,16384,0.04217955470085144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,16384,0.077565332253774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,12288,0.05622400177849663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,16384,0.04431822233729892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,12288,0.03314400050375197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,12288,0.03494844502872891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,10240,0.02907022171550327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,10240,0.046606222788492836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,65536,0.17865422036912706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,8192,0.038819554779264666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,10240,0.030768887864218816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,8192,0.028431110911899145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,7168,0.03394044438997904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,7168,0.021351110604074266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,7168,0.02275644408331977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,6144,0.029867556360032823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,6144,0.019415999452273052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,5120,0.02533333334657881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,6144,0.020720889170964558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,5120,0.017047999633683097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,4096,0.021478222476111516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,5120,0.018805333309703402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,4096,0.014730667074521383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,3584,0.01917155583699544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,4096,0.015736889508035447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,3584,0.013711111413107978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,3584,0.014646222194035849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,3072,0.01684533390733931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,3072,0.01260355528857973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,2560,0.01460977726512485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,3072,0.013337777720557319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,2560,0.011468444433477191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,2048,0.012279111478063794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,2048,0.00998488896422916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,2560,0.012375999655988483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,2048,0.011039111349317761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,1536,0.009974222216341231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,1536,0.008940444224410588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,1536,0.009842666486899057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,1024,0.007440889047251807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,1024,0.007052444749408298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,1024,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,768,0.005663110978073544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,768,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,768,0.006433777925040986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,512,0.004461333155632019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,512,0.005079999979999331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,512,0.006055999961164262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,256,0.004690666579537922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,256,0.0033200000309281876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,256,0.00564177789621883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,128,0.0030497778207063675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,128,0.00444444434510337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,64,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,128,0.005398222141795688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,64,0.005446222093370225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,6144,32,0.002903999967707528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,6144,32,0.005469333380460739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,12288,0.04053689042727152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,65536,0.22657333479987252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,16384,0.04577155576811897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,65536,0.1315822203954061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,16384,0.058507555060916476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,65536,0.13424266709221735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,10240,0.03480888737572564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,16384,0.04171289006868998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,12288,0.03044355577892727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,12288,0.04515999886724684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,12288,0.02996622191535102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,10240,0.0380951099925571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,10240,0.026639110512203638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,10240,0.026165333059098985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,8192,0.031138665146297876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,8192,0.021850667066044275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,7168,0.0277164445983039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,7168,0.020066666934225295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,6144,0.01830577850341797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,6144,0.024772443705134924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,6144,0.018146667215559218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,5120,0.016215110818545025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,5120,0.021135111649831135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,5120,0.016386666231685214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,4096,0.013942221800486246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,4096,0.017605332864655387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,4096,0.013879111243618859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,3584,0.015609777635998197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,3584,0.01293244461218516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,8192,3072,0.014379555980364481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,3584,0.01298577752378252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,3072,0.0118222220076455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,3072,0.013839999834696451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,3072,0.012125333150227865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,2560,0.012126222252845764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,2560,0.010802666346232096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,2048,0.0101742222905159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,2560,0.011200889117187925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,2048,0.010303999814722273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,1536,0.008456889126035903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,1536,0.008593777815500895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,1536,0.009306666751702627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,1024,0.006705777926577463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,1024,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,768,0.005848000033034219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,768,0.00535022219022115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,768,0.006151999864313338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,512,0.003569777641031477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,512,0.005045333256324132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,512,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,256,0.0046160001721647055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,256,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,256,0.005512889060709212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,128,0.003024000053604444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,128,0.004357333398527569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,128,0.005372444374693765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,64,0.0028728888266616394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,64,0.005425777700212266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,5120,32,0.002867555452717675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,5120,32,0.005400889035728242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,65536,0.18242133988274467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,16384,0.030172444052166406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,65536,0.11857244703504775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,16384,0.04965155654483371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,16384,0.03178044491344028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,12288,0.024316444993019104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,12288,0.03952444593111674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,10240,0.02143200072977278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,12288,0.0259226659933726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,10240,0.03350755572319031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,10240,0.02273955610063341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,8192,0.017785777648289997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,8192,0.028063111835055884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,8192,0.018945778409639995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,7168,0.01649599936273363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,7168,0.02550933261712392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,7168,0.01759111053413815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,6144,0.015223110715548197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,6144,0.022729777627521094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,6144,0.01607822212908003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,5120,0.013542222479979197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,7168,5120,0.018766222728623282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,5120,0.0176862229903539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,5120,0.014597333139843412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,4096,0.01186044431395001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,4096,0.014614222778214348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,4096,0.012531555361217923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,3584,0.011055111057228513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,3584,0.013159111142158508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,3584,0.011874666644467248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,3072,0.010252444280518426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,3072,0.011775111158688864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,3072,0.01109155515829722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,2560,0.010339555641015371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,2560,0.010416000253624385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,2048,0.008650666309727563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,2048,0.008889777792824639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,2048,0.009478222164842818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,1536,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,1536,0.007584888901975419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,1536,0.008031110796663497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,1024,0.005498666730191972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,1024,0.005984888722499211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,1024,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,768,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,768,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,768,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,512,0.004895110925038655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,512,0.003485333174467087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,512,0.005695999910434087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,256,0.004543111142185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,256,0.003148444410827425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,256,0.00546577779783143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,128,0.004293333325121137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,128,0.0029200000895394217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,128,0.005286222116814719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,64,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,64,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,4096,32,0.0029111111329661477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,4096,32,0.005305777821275923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,65536,0.09528355465994941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,65536,0.16266222794850668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,6144,8192,0.023557333482636347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,16384,0.02840177880393134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,16384,0.04459822177886963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,65536,0.09670933087666829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,16384,0.029327998558680218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,12288,0.03556888964441087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,12288,0.023056000471115112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,12288,0.02414400047726101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,10240,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,10240,0.03055466545952691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,10240,0.021470222208235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,8192,0.025669333007600572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,8192,0.01701155636045668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,8192,0.018008889423476327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,7168,0.02315555512905121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,7168,0.01590755581855774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,7168,0.01678400072786543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,6144,0.014560888210932413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,6144,0.020803555846214294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,6144,0.015298666225539314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,5120,0.012835555606418185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,5120,0.018226666582955253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,4096,0.013080000049538083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,5120,0.014024888475735983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,4096,0.011328889264000786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,4096,0.011934222446547614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,3584,0.011869333684444427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,3584,0.011364444262451597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,3072,0.009887999958462184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,3072,0.010760889285140567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,3072,0.01070844464831882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,2560,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,2560,0.009517333573765224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,2560,0.010039110978444418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,2048,0.008266666697131263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,2048,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,2048,0.009038222332795462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,1536,0.006695999867386288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,1536,0.007130666739410824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,1536,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,1024,0.005712888720962737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,1024,0.00552977787123786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,1024,0.006126222097211414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,768,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,768,0.004301333179076512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,768,0.005974222388532426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,512,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,512,0.004870222260554631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,512,0.005728888842794631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,256,0.0031351111829280853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,256,0.004466666529575984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,256,0.005468444277842839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,128,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,128,0.004209777961174647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,128,0.005292444593376583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,64,0.002728888971938027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,64,0.005335999859703912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3584,32,0.002798222212327851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3584,32,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,65536,0.07947288619147407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,7168,0.02110933264096578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,65536,0.14023289415571424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,8192,0.023597333166334365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,16384,0.041407111618253924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,16384,0.029862221744325426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,16384,0.02717777755525377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,12288,0.021465778350830078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,65536,0.08812799718644883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,12288,0.03281155559751723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,12288,0.022424888279702928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,10240,0.027858667903476294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,10240,0.019384000036451552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,10240,0.019811555743217468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,8192,0.016741333736313712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,8192,0.02405066622628106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,8192,0.016766221986876596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,7168,0.015565334094895257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,7168,0.02159999973244137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,5120,2048,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,7168,0.015643555257055495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,6144,0.019312888383865356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,6144,0.014543111125628153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,5120,0.012479999827014076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,5120,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,5120,0.013268444273206921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,4096,0.014681778020328946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,4096,0.01110400011142095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,4096,0.011443555355072021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,3584,0.010973333484596677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,3584,0.010328888893127441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,3584,0.010972444381978778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,3072,0.00982488857375251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,3072,0.010323555933104621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,2560,0.008857777549160851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,2560,0.00887022250228458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,2560,0.009647111097971598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,2048,0.007795555724038019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,2048,0.007537778053018782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,2048,0.008412444757090675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,1536,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,1536,0.006755555669466655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,1536,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,1024,0.005452444569932089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,1024,0.005131555514203178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,1024,0.006128888991143968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,768,0.0038311113086011675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,768,0.005162666655249066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,768,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,512,0.004847110973464118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,512,0.0034711110509104202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,512,0.005693333430422678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,256,0.003129777808984121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,256,0.00443733317984475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,256,0.005486222190989389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,128,0.0041440000964535605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,128,0.0029057777590221832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,64,0.0027511111564106415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,128,0.0052693333062860705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,3072,32,0.0027635554886526535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,64,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,3072,32,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,65536,0.11893599563174778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,65536,0.10048977533976237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,16384,0.035084442959891424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,65536,0.07783910963270399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,16384,0.023638221952650283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,12288,0.02730488942729102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,16384,0.025601777765485976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,12288,0.019707555572191875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,65536,0.08083911074532403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,12288,0.0207368897067176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,10240,0.023631999890009563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,10240,0.01762488815519545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,8192,0.020045333438449435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,8192,0.015099555253982544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,10240,0.018558222386572096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,8192,0.015809777710172866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,7168,0.014136888914638095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,7168,0.018210666047202218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,7168,0.014925332532988654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,6144,0.012974222501118978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,6144,0.016166221764352586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,6144,0.013722666435771518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,5120,0.014422222971916199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,5120,0.01180622229973475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,5120,0.012636444634861417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,4096,0.012443555725945367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,4096,0.010417777631017896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,4096,0.010961777634090848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,3584,0.00981155534585317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,3584,0.011662222445011139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,3584,0.010566222170988718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,3072,0.01088444474670622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,3072,0.009248889154858058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,3072,0.009884444375832876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,2560,0.0085235552655326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,2560,0.008247111406591203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,2560,0.009038222332795462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,2048,0.007139555282062954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,2048,0.0075342224703894714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,2048,0.006768888897365994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,1536,0.0058559998869895935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,1536,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,1536,0.006521777974234686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,1024,0.004528000122971005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,1024,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,1024,0.006124444305896759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,768,0.003839111162556542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,768,0.005160000175237656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,768,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,512,0.004807999978462855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,512,0.0035031110876136353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,512,0.005690666536490123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,256,0.003124444435040156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,256,0.005398222141795688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,256,0.004408888932731417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,128,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,128,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,64,0.0027217778066794076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,64,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2560,32,0.0027831111931138565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2560,32,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,65536,0.09632266892327203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,65536,0.06968622075186835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,16384,0.02156177825397915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,4096,2560,0.009539555344316695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,65536,0.0745697749985589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,16384,0.028007109959920246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,16384,0.023543111152119104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,12288,0.022720000810093347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,12288,0.017303110824690927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,12288,0.019508444600635104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,10240,0.019948444432682462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,10240,0.015471110741297403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,10240,0.017473777135213215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,8192,0.016600888636377122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,8192,0.013497778111033969
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,7168,0.015239111251301236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,8192,0.014800889624489678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,7168,0.012614222036467658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,7168,0.01402311192618476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,6144,0.013825777504179211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,6144,0.011649777491887411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,6144,0.013075555364290873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,5120,0.012525333298577202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,5120,0.010648888846238455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,5120,0.012059555285506778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,4096,0.010970667004585266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,4096,0.010427555276287926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,3584,0.009128000173303816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,3584,0.010330666270520952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,3584,0.009925333162148794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,3072,0.008114666574531132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,3072,0.009703999592198266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,3072,0.009123555488056606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,2560,0.006878222028414409
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,2560,0.007511999871995714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,2560,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,2048,0.006259555617968242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,2048,0.006690666493442323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,2048,0.007044444481531779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,1536,0.005345777918895085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,1536,0.0058151111006736755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,1536,0.006337777607970768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,1024,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,1024,0.004002666721741359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,1024,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,768,0.005063111169470681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,768,0.003719999972316954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,768,0.0057502223385704895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,512,0.0047884442740016514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,512,0.0033742222521040174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,512,0.005542222410440445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,256,0.00444533344772127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,256,0.003039999968475766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,256,0.005335111171007156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,128,0.004322666674852371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,128,0.0027946666296985415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,128,0.0052177777720822235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,64,0.0026497778793176017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,2048,32,0.0026871110830042097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,64,0.005249777601824866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,2048,32,0.005219555563396878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,65536,0.07477421893013848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,65536,0.06876799795362684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,16384,0.02092355489730835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,16384,0.02295644415749444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,16384,0.02217777735657162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,65536,0.06985866361194186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,12288,0.016880000631014507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,12288,0.01866755551762051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,12288,0.01832711034350925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,10240,0.01513066722287072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,10240,0.01645155582163069
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,8192,0.013905778527259827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,10240,0.016622222132152982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,8192,0.013191111385822296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,7168,0.013014222184816996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,8192,0.014122666584120857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,7168,0.012180444267061023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,7168,0.01348088930050532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,6144,0.011271110839313932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,6144,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,6144,0.012502222425407834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,5120,0.010992888775136737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,5120,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,5120,0.01165244479974111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,4096,0.009879999690585667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,4096,0.009138666921191746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,4096,0.009920888476901585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,3584,0.008172444171375698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,3584,0.009355555805895064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,3584,0.009034666750166152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,3072,0.007248889240953658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,3072,0.008908444808589088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,3072,0.008057778080304464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,2560,0.006716444260544247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,2560,0.007069333559936947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,2560,0.007247111035717859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,2048,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,2048,0.006231110956933763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,2048,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,1536,0.0058106668293476105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,1536,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,1536,0.006533333410819371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,1024,0.005323555734422471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,1024,0.004058666527271271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,1024,0.005979555762476391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,768,0.005041777673694822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,768,0.003660444584157732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,768,0.0058133333093590206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,512,0.0047582222355736625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,512,0.0033617777129014335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,256,0.004410666724046071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,256,0.0030088888274298776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,256,0.005377777748637729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1536,128,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,128,0.0027875554644399216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,128,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,64,0.0028231110837724474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,64,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1536,32,0.0026862221873468826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1536,32,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,65536,0.05316177672810025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,65536,0.06783733103010389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,16384,0.020603555772039626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,65536,0.06664178106519911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,16384,0.017180444465743173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,16384,0.02128711177243127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,12288,0.014177777700954013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,12288,0.017621333400408428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,10240,0.014849777022997538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,10240,0.01278133359220293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,10240,0.0159608887301551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3584,3584,0.010609777437316047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,8192,0.012842666771676807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,8192,0.011203555597199334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,8192,0.013723555538389417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,7168,0.010568888651000129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,7168,0.01294222225745519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,6144,0.010627555350462595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,6144,0.009583111438486312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,6144,0.011881777809725868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,5120,0.009202666580677032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,5120,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,5120,0.010128888818952773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,4096,0.008023999631404877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,4096,0.008608889248636033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,4096,0.008839111361238692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,3584,0.007522666619883643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,3584,0.008392888638708327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,3584,0.007954667011896769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,3072,0.007019555403126611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,3072,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,3072,0.007524444825119442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,2560,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,2560,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,2560,0.007176000210973952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,2048,0.006184000107977126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,2048,0.005281777845488654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,2048,0.006702222343948152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,1536,0.005709333138333426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,1536,0.004631999880075455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,1536,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,1024,0.003941333128346337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,1024,0.0053119998839166425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,1024,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,768,0.0036195553839206696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,768,0.005050666630268097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,768,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,512,0.0047155556579430895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,512,0.003302222324742211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,512,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,256,0.004323555363549126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,256,0.0029928889125585556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,256,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,128,0.0027644443843099806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,128,0.004303110970391167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,128,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,64,0.0026480000880029467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,64,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,1024,32,0.0026382222357723448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,1024,32,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,65536,0.06843911276923285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,65536,0.042968001630571157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,65536,0.06688355737262301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,16384,0.020560888780487906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,16384,0.014496889379289416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,16384,0.021253334151373968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,12288,0.01651822196112739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,12288,0.01239733315176434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,12288,0.017567111386193168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,3072,0.009576000273227692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,10240,0.014726221561431885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,10240,0.0114702218108707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,3072,6144,0.01422933406300015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,10240,0.015936000479592215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,8192,0.01018133345577452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,8192,0.013584000368913015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,7168,0.009485333330101436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,7168,0.01257777793539895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,6144,0.010658666491508484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,6144,0.008763555851247575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,6144,0.011471110913488599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,5120,0.009003555609120263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,5120,0.008607111043400234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,5120,0.009504000345865885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,4096,0.008017777568764156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,4096,0.0075502221783002215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,4096,0.008454221818182204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,3584,0.007528000407748752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,3584,0.007463999920421177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,3584,0.008017777568764156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,3072,0.007053333024183909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,3072,0.00721777809990777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,3072,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,2560,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,2560,0.006165333506133821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,2560,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,2048,0.00620888877246115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,2048,0.005361777626805835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,2048,0.006794666664467917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,1536,0.005720888988839255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,1536,0.0046897778908411665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,1536,0.0063733334342638654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,1024,0.005304000029961268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,1024,0.003962666624122196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,1024,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,768,0.004970666848950916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,768,0.0036142224238978494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,768,0.0057911111248864075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,512,0.0046773333516385825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,512,0.0032924444725116095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,512,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,256,0.004291555533806483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,256,0.002966222249799304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,256,0.005372444374693765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,128,0.004127999974621667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,128,0.0028000000036425064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,128,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,64,0.0026311110705137253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,64,0.005201777650250329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,768,32,0.0026213334252436957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,768,32,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,65536,0.03167377909024557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,65536,0.06777955426110162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2560,128,0.0042160000238153665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,16384,0.011552888486120435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,65536,0.06554310851626925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,16384,0.02051377793153127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,16384,0.021197333931922913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,12288,0.010112000008424124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,12288,0.01611822181277805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,12288,0.017320889565679762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,10240,0.009407111340098912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,10240,0.014089778065681458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,10240,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,8192,0.011870221959220039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,8192,0.008823110825485652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,8192,0.012166221936543783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,7168,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,7168,0.010879110958841113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,6144,0.007453333172533248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,7168,0.011295999917719098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,6144,0.009933333430025313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,6144,0.01019022199842665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,5120,0.007510222494602203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,5120,0.008967111508051554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,5120,0.009395555489593083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,4096,0.00639111093348927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,4096,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,4096,0.008341333104504479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,3584,0.006816000160243776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,3584,0.007496000164084964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,3584,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,3072,0.0069511110583941145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,3072,0.006666666517655055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,3072,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,2560,0.006017777654859755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,2560,0.006535999890830781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,2560,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,2048,0.006152888966931238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,2048,0.005346666607591841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,2048,0.006709333509206772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,1536,0.005746666755941179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,1536,0.004631999880075455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,1536,0.00638488887084855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,1024,0.005222222043408289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,1024,0.003990222182538775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,1024,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,768,0.004978666702906291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,768,0.0035991109907627106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,768,0.005791999813583162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,512,0.004643555730581284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,512,0.0033297776761982176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,512,0.005528000079923206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,256,0.004308444344335132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,256,0.002954666606254048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,256,0.005366222312053044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,512,128,0.004132444245947732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,128,0.002784888984428512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,128,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,64,0.0026488889836602737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,512,32,0.002640888922744327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,64,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,512,32,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,65536,0.021132444341977436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,65536,0.0671244462331136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,65536,0.06567466921276517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,16384,0.019467555814319186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,16384,0.008963555925422246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,16384,0.019894222418467205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,2048,4096,0.009566222627957663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,12288,0.008088889221350351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,12288,0.01538755496342977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,12288,0.015782222151756287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,10240,0.007438222567240397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,10240,0.013546666337384118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,8192,0.006873777757088344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,10240,0.013800000150998434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,8192,0.011708444191349877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,7168,0.010773333410422007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,7168,0.006524444454246097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,7168,0.010943999720944298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,6144,0.0097884444726838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,6144,0.006521777974234686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,6144,0.010183999935785929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,5120,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,5120,0.008868444297048781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,5120,0.009360888765917884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,4096,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,4096,0.007864889171388414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,4096,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,3584,0.006800000038411882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,3584,0.007927999728255803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,3072,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,3072,0.0064897777305708984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,3072,0.007511999871995714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,2560,0.0064560001095136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,2560,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,2560,0.007132444116804335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,2048,0.006052444378534953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,2048,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,2048,0.006680000159475539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,1536,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,1536,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,1536,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,1024,0.005184000151024925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,1024,0.003952888978852166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,1024,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,768,0.004920888692140579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,768,0.00360622215602133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,768,0.005708444449636672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,512,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,512,0.003302222324742211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,512,0.005472888963090048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,256,0.004243555582231945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,256,0.002938666691382726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,256,0.0052871112194326185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,128,0.004081777814361784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,128,0.0027182222240500977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,128,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,64,0.0026391111314296722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,64,0.00518844442235099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,256,32,0.002593777866827117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,256,32,0.005160000175237656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,65536,0.015042665931913586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,65536,0.06508266925811768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,16384,0.019139554765489366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,16384,0.006721777634488211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,16384,0.018486221631368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,12288,0.015258666541841296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,12288,0.0064284445510970215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,12288,0.015086222026083203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,10240,0.01350400017367469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,10240,0.006524444454246097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,10240,0.01350400017367469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,12288,0.016466667254765827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,1024,7168,0.011968889170222811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,8192,0.006416000011894438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,8192,0.011706666813956367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,7168,0.0063484443558586975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,7168,0.010994666980372535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,6144,0.009829333258999718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,6144,0.006243555496136348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,6144,0.01015466699997584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,5120,0.008867555194430882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,5120,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,5120,0.00943377779589759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,4096,0.007832888927724626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,4096,0.006171555568774541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,4096,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,3584,0.0073928890956772705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,3584,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,3584,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,3072,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,3072,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,3072,0.0075288886825243635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,2560,0.0064382221963670515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,2560,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,2560,0.007158222297827403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,2048,0.006046222315894232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,2048,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,2048,0.006698666761318843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,1536,0.005594666633341048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,1536,0.00461422238085005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,1536,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,1024,0.005143110950787862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,1024,0.003928889003064898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,1024,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,768,0.004874666531880696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,768,0.003616888903909259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,768,0.005748444547255834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,512,0.004520888957712385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,512,0.0032951111594835916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,512,0.005514666852023866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,256,0.00423555572827657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,256,0.0029288888391521242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,256,0.0053084443012873335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,128,0.004027555386225383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,128,0.0027439999911520216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,128,0.005156444592608346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,64,0.002577777744995223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,64,0.005219555563396878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,128,32,0.0026231110095977783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,128,32,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,65536,0.01347555551264021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,65536,0.0656133360332913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,16384,0.006486222147941589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,16384,0.01865600049495697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,12288,0.0063884444534778595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,12288,0.01520088811715444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,10240,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,10240,0.013600000076823764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,8192,0.006413333531883027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,8192,0.011811555259757571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,7168,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,7168,0.010996444357766045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,6144,0.006118222243256039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,6144,0.01018044435315662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,5120,0.0064444442590077715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,5120,0.009415999882751042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,4096,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,4096,0.008381333616044786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,8192,0.012664000193277994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,3584,0.006541333264774746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,3584,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,3072,0.006198222024573221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,3072,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,2560,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,2048,0.0052693333062860705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,2560,0.007383999725182851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,2048,0.006734222173690796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,1536,0.004613333278232151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,1536,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,1024,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,1024,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,768,0.0037751110891501107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,768,0.00572622236278322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,512,0.0032222221295038858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,512,0.005512889060709212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,256,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,256,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,128,0.0027182222240500977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,128,0.005161777966552311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,64,0.002570666579736604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,768,7168,0.011156444748242697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,64,0.005124444348944558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,64,32,0.002557333351837264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,64,32,0.005153777698675792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,65536,0.012297777665985955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,16384,0.006770666688680649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,16384,0.018509333332379658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,65536,0.06601955493291219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,12288,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,12288,0.015084443820847405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,10240,0.013540444274743399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,10240,0.00647199981742435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,8192,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,8192,0.011778666741318174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,7168,0.006095110956165526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,7168,0.011021333436171213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,6144,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,6144,0.010214221974213919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,5120,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,4096,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,5120,0.00942311104800966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,3584,0.006457777900828256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,4096,0.008325333396593729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,3584,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,3072,0.006048888795905643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,3072,0.00757244477669398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,2560,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,2048,0.005209777918126848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,2560,0.007180444068378872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,1536,0.004574222283230888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,2048,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,1536,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,1024,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,768,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,768,0.005736888696750005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,512,0.0032177778581778207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,512,0.005498666730191972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,256,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,128,0.0027031112048361036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,256,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,128,0.005133333305517833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,64,0.0025395556456512874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,64,0.005111999809741974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,32,32,32,0.002558222247494592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,32,32,32,0.005127999931573868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,16384,0.7149546411302355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,16384,0.3755937682257758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,12288,0.543698681725396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,16384,0.40770933363172746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,12288,0.28591820928785533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,8192,0.011788444386588203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,256,3584,0.00739911115831799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,10240,0.4572319984436035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,10240,0.26047200626797146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,12288,0.3287377887301975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,8192,0.363571564356486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,10240,0.2538684474097358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,8192,0.1799395614200168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,7168,0.31520266003078884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,7168,0.18064799573686388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,8192,0.20206844806671143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,6144,0.2717404365539551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,7168,0.1760257747438219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,6144,0.1360026730431451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,5120,0.11680800384945339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,6144,0.15376178423563638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,5120,0.23292001088460287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,4096,0.0942222211096022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,4096,0.18488889270358613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,5120,0.1406044430202908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,3584,0.08292177650663587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,3584,0.160316440794203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,4096,0.10308533244662815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,65536,0.0674826635254754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,3072,0.13829421997070312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,3584,0.09902933571073745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,3072,0.07342222001817492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,2560,0.11584177282121445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,3072,0.07689155472649468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,2560,0.06576444705327351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,2048,0.09280888901816474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,1536,0.07157155540254381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,2048,0.05116888880729675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,2560,0.0654239985677931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,2048,0.05608977874120077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,1024,0.048786666658189565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,1536,0.042846222718556724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,1024,0.031176888280444678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,768,0.03827022181616889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,768,0.023174222972657945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,768,0.02595022155178918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,512,0.02639644510216183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,512,0.02052622205681271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,256,0.015007999208238391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,256,0.012803555362754397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,256,0.016875555117925007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,128,0.009711999860074785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,128,0.008985777695973715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,64,0.0063546668324205614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,128,0.014144000079896716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,65536,32,0.00453955555955569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,64,0.01090755561987559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,65536,32,0.010971555279360877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,7168,0.010782221953074137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,32,128,8192,0.011690666278203329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,65536,0.7035759819878472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,16384,0.1743333339691162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,16384,0.09490311145782471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,12288,0.13266755474938288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,16384,0.12185066276126438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,12288,0.07248710923724704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,65536,0.3739608923594157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,12288,0.08726577626334296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,65536,0.43712888823615176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,10240,0.11835022767384847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,8192,0.052490668164359204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,10240,0.06968000200059679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,8192,0.09575644466612075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,7168,0.08447377549277411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,8192,0.059346669250064425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,7168,0.0537377761469947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,6144,0.044469333357281156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,6144,0.07836088869306776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,5120,0.03400533397992452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,6144,0.04464977648523119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,5120,0.05720000134574043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,5120,0.03847377830081516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,4096,0.028091556496090356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,4096,0.04678844412167867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,4096,0.03162489003605313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,3584,0.04146577914555868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,3584,0.02572533322705163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,3584,0.028544889556037054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,3072,0.022808000445365906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,3072,0.03606222073237101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,3072,0.025368889172871906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,2560,0.019918221566412184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,2560,0.030795554320017498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,2560,0.022312000393867493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,2048,0.016917333006858826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,2048,0.025238222546047632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,2048,0.019098665979173448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,1536,0.014162666267818876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,1536,0.019935111204783123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,1536,0.015787555111779105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,1024,0.014039999908871122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,1024,0.01130488928821352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,768,0.011188444164064197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,1024,0.012979555461141797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,768,0.009836444424258338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,768,0.011546666423479715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,512,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,512,0.008412444757090675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,512,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,256,0.005691555639108022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,256,0.004919111314747068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,256,0.007155555817815993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,128,0.0033902221669753394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,128,0.004513777792453766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,64,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,128,0.006772444479995304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,16384,32,0.0032008888406885993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,64,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,16384,32,0.006791111081838608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,65536,0.5369715690612793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,16384,0.0844951139556037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,65536,0.2834080060323079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,65536,0.3121235635545519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,16384,0.13835733466678196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,12288,0.10035733381907146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,12288,0.05590133534537422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,16384,0.08414844671885173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,10240,0.0842026670773824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,12288,0.05952622493108114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,1536,0.04001244571473863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,10240,0.047192888127432935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,8192,0.07548266649246216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,8192,0.03945066531499227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,10240,0.05157688922352261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,512,0.019504888190163504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,8192,0.04448533389303419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,7168,0.059975114133622914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,7168,0.03481866584883796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,6144,0.05272444420390659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,7168,0.037681778271993004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,65536,1024,0.03136444422933791
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,5120,0.04515555500984192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,6144,0.03372088736957974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,5120,0.026583999395370483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,4096,0.03672088848219977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,5120,0.029149333635965984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,3584,0.032644444041781955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,4096,0.02398577829202016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,3584,0.020398222737842135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,3072,0.028479999966091577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,3584,0.02219555609756046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,3072,0.01811200049188402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,3072,0.019740444090631273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,2560,0.024047111471494038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,2560,0.01609422266483307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,2560,0.017470222380426195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,2048,0.02015822297996945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,2048,0.013887999786270989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,2048,0.015084443820847405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,1536,0.011925333076053195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,1536,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,1536,0.013134222063753339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,1024,0.009689778089523315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,1024,0.01146488885084788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,1024,0.011158222125636207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,768,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,768,0.009591110878520543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,768,0.010127999716334874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,512,0.0074337778819931885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,512,0.007882666256692674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,256,0.0043333334227403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,256,0.004709333181381226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,256,0.0064026667839950984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,128,0.004374222209056218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,128,0.003391111062632667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,128,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,64,0.003085333233078321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,12288,32,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,64,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,12288,32,0.006153777655627992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,65536,0.4531688690185547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,65536,0.24206754896375868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,10240,0.060699555608961314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,16384,0.06739199823803373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,16384,0.11284977859920925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,16384,7168,0.04461866617202759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,12288,0.0861848857667711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,16384,0.06806577576531304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,65536,0.29622666041056317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,12288,0.05291733476850721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,10240,0.07237778107325236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,10240,0.041887111133999295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,8192,0.058415108256869845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,10240,0.045371555619769625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,8192,0.03425066669782003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,8192,0.03717155588997735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,7168,0.051607112089792885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,7168,0.031126221021016438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,7168,0.03380089004834493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,6144,0.04513955447408888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,5120,0.0380675560898251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,6144,0.02975377771589491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,5120,0.026053332620196875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,4096,0.019865777757432725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,4096,0.030993776188956365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,4096,0.02183466653029124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,3584,0.01843644513024224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,3584,0.027639999985694885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,3584,0.01990133358372582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,3072,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,3072,0.02423200011253357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,3072,0.01787911189926995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,2560,0.014863999353514777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,2560,0.020649777518378366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,2560,0.015942222542232938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,2048,0.012851555314328937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,2048,0.017097777790493436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,2048,0.013931555880440606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,1536,0.011063111325105032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,1536,0.013394667042626275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,1536,0.012284444438086616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,1024,0.009200000100665623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,1024,0.009938666390048133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,1024,0.010575110713640848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,768,0.008264000217119852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,768,0.008235555556085374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,768,0.009743111001120673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,512,0.005375111268626318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,512,0.006520000182920032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,512,0.006999111009968652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,256,0.0048177776237328844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,256,0.0034888887570963967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,256,0.0063386667105886675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,128,0.004494222087992562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,128,0.0031315556002987754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,128,0.00602311102880372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,64,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,64,0.006026666611433029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,10240,32,0.0029048888633648553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,10240,32,0.0060159998635451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,65536,0.3557751178741455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,65536,0.1949359973271688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,16384,0.05183644427193535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,16384,0.08880888753467137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,65536,0.26501332388983834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,16384,0.060702224572499595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,12288,0.040004445446862116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,12288,0.06791822115580241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,12288,0.04731288883421156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,10240,0.034323556555642024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,10240,0.05686399671766493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,10240,0.041187554597854614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,8192,0.028425776296191748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,8192,0.04628711276584201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,8192,0.03368533319897122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,7168,0.02546666728125678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,7168,0.04095110959476895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,7168,0.030212445391549006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,6144,0.02276977731121911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,6144,0.03587911195225186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,6144,0.02684355609946781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,5120,0.030677331818474665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,5120,0.01997866729895274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,6144,0.030870222383075293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,5120,0.02363377809524536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,4096,0.017078222499953378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,4096,0.02513155506716834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,4096,0.02014133334159851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,3584,0.022355554832352534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,3584,0.018373333745532565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,4096,0.022397332721286353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,3072,0.014275555809338888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,3072,0.01982399986849891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,12288,512,0.0069466663731469055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,3072,0.016382222374280293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,2560,0.016931555337376065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,2560,0.015038222074508667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,2048,0.011109333071443768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,2048,0.013929777675204806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,2048,0.013036444783210754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,1536,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,1536,0.011219555305110084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,1536,0.011763555308183035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,1024,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,1024,0.008449777960777283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,1024,0.009883555273214975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,768,0.007132444116804335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,768,0.008344888687133789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,512,0.005113777601056629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,512,0.005783111270931032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,512,0.006720888945791457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,256,0.004590222405062782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,256,0.0034479999707804788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,256,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,128,0.004287999951177173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,128,0.003050666716363695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,128,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,64,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,64,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,8192,32,0.0029599999801980127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,8192,32,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,65536,0.17754666010538736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,12288,0.04884533418549431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,65536,0.30690399805704754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,16384,0.09512622488869561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,16384,0.04662044511901008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,16384,0.04956000049908956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,12288,0.037235554721620344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,12288,0.06420355372958712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,5120,0.024171556035677593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,12288,0.038555555873446994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,65536,0.1875173383288913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,10240,0.032101333141326904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,10240,0.05436977744102478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,10240,6144,0.02757511039574941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,8192,0.04148533278041416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,10240,0.033253331979115806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,8192,0.029416001505321924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,7168,0.03664977682961358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,8192,0.02810488806830512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,7168,0.02546666728125678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,6144,0.032362666394975446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,6144,0.021307556165589228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,5120,0.02754222187731001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,6144,0.0229768885506524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,5120,0.018752889500723947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,5120,0.02033511135313246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,4096,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,4096,0.02271911170747545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,4096,0.01682044400109185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,3584,0.014721777704026965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,3584,0.0203484445810318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,3584,0.01565511193540361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,3072,0.018233777748213876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,3072,0.013426666458447775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,3072,0.014254222313563028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,2560,0.015533333023389181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,2560,0.01201866649919086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,2560,0.01314755529165268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,2048,0.01295733369059033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,2048,0.0105262224872907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,2048,0.011624000138706632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,1536,0.010503999888896942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,1536,0.009333333207501305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,1536,0.010245333115259806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,1024,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,1024,0.00885688927438524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,768,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,768,0.006778666542636023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,768,0.0071937781241205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,512,0.0053324442770746015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,512,0.00499911109606425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,512,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,256,0.0032702222880389956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,256,0.004610666798220741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,256,0.005713777823580637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,128,0.003047111133734385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,128,0.005452444569932089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,128,0.0042204442951414315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,64,0.002893333426780171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,64,0.005477333234416113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,7168,32,0.0028479999552170434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,7168,32,0.005538666827811136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,65536,0.26768265830145943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,16384,0.04720977942148844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,65536,0.14875555038452148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,65536,0.1597555610868666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,16384,0.07333155473073323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,16384,0.046711110406451754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,12288,0.03288177649180094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,12288,0.0551528897550371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,12288,0.03462933169470893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,10240,0.028208888239330713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,10240,0.0458568897512224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,10240,0.029503110382292006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,8192,0.023173333870040044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,8192,0.03814844290415446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,8192,0.02535377773973677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,7168,0.021176889538764954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,7168,0.033477332856920027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,7168,0.022712000542216833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,6144,0.029464887248145208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,6144,0.01925155520439148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,6144,0.0210826661851671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,5120,0.016912889149453905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,5120,0.02495999965402815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,5120,0.018598222070270114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,4096,0.014495111174053617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,4096,0.021112889051437378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,4096,0.015523556205961438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,3584,0.013440888788965015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,3584,0.01885333326127794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,3072,0.012334222594896952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,3584,0.014682667122946845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,3072,0.016562667157914903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,3072,0.013530666629473368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,2560,0.014434667097197639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,2560,0.011261333194043903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,2560,0.012296888563368054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,2048,0.009999111294746399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,2048,0.012168889244397482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,2048,0.010851555400424533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,1536,0.00980533328321245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,1536,0.008861333131790161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,1536,0.00978133330742518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,1024,0.00741155528359943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,1024,0.00702133360836241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,1024,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,768,0.005516444643338521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,768,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,768,0.006542222367392645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,512,0.005056000004212062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,512,0.0040755557517210645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,512,0.006048000107208888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,256,0.0032248888164758682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,256,0.0046817776229646476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,256,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,128,0.003010666618744532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,6144,128,0.004359999878538979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,64,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,128,0.005428444594144821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,6144,32,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,64,0.005390222287840313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,6144,32,0.005406222409672207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,3584,0.015767110718621146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,65536,0.22340001000298393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,65536,0.13188889291551378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,16384,0.046112000942230225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,16384,0.05827555391523573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,12288,0.044587555858823985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,16384,0.03738311264250014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,2560,0.012878222597969903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,10240,0.026537777649031744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,65536,0.13169688648647732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,12288,0.029672000143263075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,10240,0.037717332442601524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,8192,0.030771556827757094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,8192,768,0.0068897778789202375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,10240,0.02586222191651662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,8192,0.02193244463867611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,7168,0.02753955622514089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,8192,0.02178222272131178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,7168,0.02106133268939124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,7168,0.02001333402262794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,6144,0.024280000064108107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,6144,0.01791200041770935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,6144,0.018168000711335074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,5120,0.02071555621094174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,5120,0.01607111096382141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,5120,0.016139555308553908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,4096,0.017284444636768766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,4096,0.013735999663670858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,4096,0.013834666874673633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,3584,0.015413333972295126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,3584,0.012672888735930124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,3072,0.013603555659453074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,3584,0.013097777962684631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,3072,0.011681777735551199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,3072,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,2560,0.011896889242861005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,2560,0.010641777680979835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,2560,0.011247110863526663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,2048,0.010176888770527309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,2048,0.009519110951158736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,2048,0.01013777818944719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,1536,0.008550222549173567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,1536,0.00849599970711602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,1536,0.009297777381208207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,1024,0.006671111202902264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,1024,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,1024,0.007213333414660559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,768,0.005389333185222413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,768,0.0058515556156635284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,768,0.006174222048785951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,512,0.0035911111368073356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,512,0.004971555537647671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,512,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,256,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,256,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,256,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,128,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,128,0.004307555655638377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,128,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,64,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,64,0.005336888962321811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,5120,32,0.0027742222365405825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,5120,32,0.005341333233647876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,16384,0.029743999242782593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,16384,0.04901599884033203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,65536,0.17749244636959502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,16384,0.03151111139191522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,12288,0.02674399978584713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,65536,0.09984178013271755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,65536,0.1074977782037523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,12288,0.039064887497160174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,12288,0.02583733366595374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,10240,0.03315733207596673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,10240,0.022480888499153986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,10240,0.021367111139827307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,8192,0.02760444415940179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,7168,0.01625955601533254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,8192,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,7168,0.02500533395343357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,6144,0.022329777479171753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,7168,0.01738755570517646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,6144,0.015106666419241162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,6144,0.01605244477589925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,5120,0.019279999865425956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,5120,0.014527110589875115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,4096,0.011671110987663269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,4096,0.016357334123717416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,4096,0.01236533373594284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,3584,0.01479822231663598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,3584,0.011065777805116443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,3584,0.011711111499203576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,3072,0.010228444304731157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,3072,0.011685333318180509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,3072,0.011033777561452655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,2560,0.009448889229032729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,2560,0.010251555177900527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,2560,0.010232888989978367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,2048,0.00852711084816191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,2048,0.008755555583371056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,2048,0.009475555684831407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,1536,0.006846222198671765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,1536,0.0075191110372543335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,1536,0.007466666400432587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,1024,0.005440000030729506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,1024,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,1024,0.006132444573773279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,768,0.005124444348944558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,768,0.004593777573770947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,768,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,512,0.004807999978462855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,512,0.0034560000316964257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,512,0.005673777725961473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,256,0.00313688897424274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,256,0.004494222087992562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,256,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,128,0.00425866660144594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,128,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,128,0.005205333232879639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,64,0.0027439999911520216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,64,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,4096,32,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,4096,32,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,65536,0.15812444686889648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,65536,0.09205689032872517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,65536,0.10461688703960842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,7168,0.023966223001480103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,16384,0.044008000029457935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,16384,0.02901155418819851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,7168,1024,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,12288,0.022617777188618977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,12288,0.03444622291458978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,12288,0.02386844489309523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,10240,0.02016799979739719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,10240,0.029789331886503432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,10240,0.0210524449745814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,8192,0.016768000192112394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,8192,0.025264889001846313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,7168,0.02276444435119629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,8192,0.017487999465730455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,7168,0.015509333875444202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,6144,0.020429333051045735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,7168,0.016424889365832012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,6144,0.01426488823360867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,6144,0.015087111128701104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,5120,0.012727999852763282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,5120,0.01371022231049008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,5120,0.017669333351982962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,4096,0.011182222101423474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,4096,0.01185955521133211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,4096,0.0151235560576121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,3584,0.01364177796575758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,3584,0.011228444675604502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,3584,0.010558221903112201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,3072,0.010690666735172272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,3072,0.01056177748574151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,2560,0.009474666582213508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,2560,0.009147555463843876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,2560,0.009937778115272522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,2048,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,2048,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,2048,0.009113777842786577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,1536,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,1536,0.007118222614129384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,1536,0.006881777611043718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,1024,0.005766222046481238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,1024,0.005477333234416113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,1024,0.006117333140638139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,768,0.00387999994887246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,768,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,512,0.004853333450025982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,512,0.003458666718668408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,512,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,256,0.004457777573002709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,256,0.003121777748068174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,256,0.005414222263627582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,128,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,128,0.002830222249031067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,128,0.005255110975768831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,64,0.0027111110587914786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,64,0.005286222116814719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3584,32,0.002757333426011933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3584,32,0.0053022222386466126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,65536,0.0779706637064616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,65536,0.13915733496348062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,65536,0.09321066406038071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,16384,0.03933066791958279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,16384,0.026889777845806543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,12288,0.021811554829279583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,12288,0.030847999784681533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,12288,0.0221733334991667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,10240,0.019646222392718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,10240,0.027122666438420612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,10240,0.019525332583321463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,8192,0.016571554872724745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,8192,0.023431110713216994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,8192,0.01649688846535153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,7168,0.015290665957662793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,7168,0.02110933264096578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,5120,12288,0.032093332873450384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,7168,0.015446222490734525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,6144,0.018723555737071566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,6144,0.014154665999942355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,6144,0.014167111780908374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,5120,0.012510221865442065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,5120,0.016407999727461074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,5120,0.013088000317414602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,4096,0.010844444235165914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,4096,0.014379555980364481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,4096,0.011335111326641507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,3584,0.010105777945783403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,3584,0.01314399970902337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,3584,0.01076533314254549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,3072,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,3072,0.01198755535814497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,3072,0.010120000276300643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,2560,0.008843555218643611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,2560,0.008824889030721452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,2560,0.009470221896966299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,2048,0.007215110792054071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,2048,0.007660444411966536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,2048,0.007877333296669854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,1536,0.006725333217117522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,1536,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,1536,0.006568000134494569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,1024,0.005485333502292633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,1024,0.004878222114510006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,1024,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,768,0.005111111121045219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,768,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,768,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,512,0.003452444449067116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,512,0.004815111143721474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,512,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,256,0.00309599998096625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,256,0.005404444618357553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,128,0.004173333446184794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,128,0.0028595555987623003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,128,0.0052168890833854675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,64,0.0026942222482628296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,64,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,3072,32,0.0027093332674768236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,8192,0.017817777064111497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,4096,5120,0.013366222381591797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,3072,32,0.0053297777970631914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,65536,0.11595022678375244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,65536,0.07638133234447904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,16384,0.023716444770495098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,16384,0.03328088919321696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,16384,0.024912888805071514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,65536,0.08137689034144084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,12288,0.01980622278319465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,12288,0.026776888304286536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,12288,0.02051999999417199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,10240,0.02314666741424137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,10240,0.018205333087179396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,8192,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,8192,0.019545776976479422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,8192,0.015543111496501498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,7168,0.013968888256284924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,7168,0.017839110559887357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,7168,0.01456622282663981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,6144,0.01584533353646596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,6144,0.013023110727469126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,6144,0.01350044459104538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,5120,0.011665778027640449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,5120,0.014025777578353882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,5120,0.012475555141766867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,4096,0.010265777508417765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,4096,0.012346666720178394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,4096,0.010840000377760993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,3584,0.009759999811649323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,3584,0.011557333171367645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,3584,0.010449777874681683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,16384,0.028255999088287354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,3072,0.009153777526484596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,3072,0.010729778144094678
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,3072,0.00980088859796524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,2560,0.008184889124499427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,2560,0.008929777476522658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,2048,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,2048,0.007129777636792924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,2048,0.0068711112770769335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,1536,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,1536,0.005856888989607493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,1536,0.0064693333374129395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,1024,0.00545155546731419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,1024,0.004553777890072929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,1024,0.006070222291681502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,768,0.003801777958869934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,768,0.005104889058404499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,768,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,512,0.004753777964247597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,512,0.0034257777863078644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,512,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,256,0.004401777767472797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,256,0.003075555587808291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,256,0.005414222263627582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,128,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,128,0.004120888809363048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,64,0.002691555561290847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,128,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2560,32,0.0027502222607533136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,64,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2560,32,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,65536,0.09271377987331814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,3072,0.009822222093741098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,65536,0.06854489114549425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,16384,0.027464000715149775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,16384,0.023294222023752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,65536,0.07364444600211249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,12288,0.01720977822939555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,12288,0.022269333402315777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,12288,0.019107555349667866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,10240,0.015251555376582675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,10240,0.019461333751678467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,10240,0.01719377769364251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,8192,0.013352000051074557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,8192,0.01626133256488376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,8192,0.014601776997248331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,7168,0.012438221938080259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,7168,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,7168,0.013782222237851886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3584,768,0.005131555514203178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,6144,0.013444444371594323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,6144,0.011515555282433828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,5120,0.010590222146775989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,6144,0.01292711082432005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,5120,0.012295111185974546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,5120,0.011935999823941125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,4096,0.0107351111041175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,4096,0.00960000024901496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,4096,0.010276444256305695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,3584,0.008991111483838823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,3584,0.009857777920034196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,3584,0.010216000179449717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,3072,0.009538667069541084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,3072,0.007862221863534715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,3072,0.009124444590674506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,2560,0.006683555742104848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,2560,0.007524444825119442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,2560,0.0075253330998950535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,2048,0.006670222100284364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,2048,0.006229333165619109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,2048,0.006913777854707506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,1536,0.005741333381997214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,1536,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,1536,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,1024,0.003993777765168084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,1024,0.005296000176005893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,1024,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,768,0.005028444445795483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,768,0.003670222229427762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,768,0.005745777653323279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,512,0.0033208889265855155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,512,0.004707555390066571
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,512,0.005535999933878581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,256,0.004422222160630756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,256,0.00299822228650252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,256,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,128,0.0027964444210131965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,128,0.004275555411974589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,128,0.005160000175237656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,64,0.0026293332791990707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,64,0.005138666679461797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,2048,32,0.0026666666898462507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,2048,32,0.005142222262091107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,65536,0.07205244567659166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,65536,0.06722933053970337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,16384,0.020621332857343886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,16384,0.022593778040673997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,65536,0.06888088915083143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,16384,0.0218577782313029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,12288,0.0166302224000295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,12288,0.01834222177664439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,12288,0.01808355583084954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,10240,0.016008888681729633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,10240,0.01479644411140018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,10240,0.01629155543115404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,8192,0.012806221842765808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,8192,0.013783111340469785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,16384,0.02632444434695774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,8192,0.013905778527259827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,7168,0.012877333495352002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,7168,0.01202311118443807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,7168,0.01315199997689989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,6144,0.011767110890812345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,6144,0.011180444724029966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,6144,0.012313777373896705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,5120,0.010255999863147736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,5120,0.01089333328935835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,5120,0.011353777514563667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,4096,0.009717333647939894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,4096,0.008601778083377415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,4096,0.009787555370065901
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,3584,0.009215999808576372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,3584,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,3072,0.007032000356250339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,3072,0.008746667040718926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,3072,0.0075528886583116316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,2560,0.00655644428398874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,2560,0.007096889118353526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,2560,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,2048,0.006178666734033161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,2048,0.006093333164850871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,2048,0.006714666469229593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,1536,0.0057297779454125305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,1536,0.004680888934267892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,1536,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,1024,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,1024,0.003960888832807541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,1024,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,768,0.003630222131808599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,768,0.004952888935804367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,768,0.005728000154097875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,512,0.004690666579537922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,512,0.0032986667421129015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,512,0.005524444497293896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,256,0.00434755575325754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,256,0.002955555501911375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,256,0.005527110977305307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,128,0.004106666478845808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,128,0.0027502222607533136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,128,0.005192888693677055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,64,0.0026364444444576898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1536,32,0.002612444468670421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,64,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1536,32,0.00519466648499171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,65536,0.04914399981498718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,65536,0.06638755400975545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,16384,0.02016266683737437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,65536,0.06591111421585083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,16384,0.01682577696111467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,16384,0.021047110358874004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,12288,0.01631733278433482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,12288,0.01400888959566752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,12288,0.017415111263593037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,10240,0.012518222133318583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,10240,0.014497778481907315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,10240,0.015827554795477126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,8192,0.012568888564904531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,8192,0.011013333168294696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,8192,0.013603555659453074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,7168,0.010406222608354356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,7168,0.011635555161370171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,7168,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,6144,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,6144,0.009464888936943477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,6144,0.011508444117175208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,5120,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,5120,0.00908622228437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,5120,0.010124444133705562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,4096,0.008401778009202745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,4096,0.007902222375075022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,4096,0.008291555775536431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,3584,0.007450666692521837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,3584,0.008099555969238281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,3584,0.007920888562997183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,3072,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,3072,0.007167110840479533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,3072,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,2560,0.0064995553758409284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,2560,0.006481777876615524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,2560,0.007136888802051544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,2048,0.006144000010357962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,2048,0.005370666583379109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,2048,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,1536,0.005664888769388199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,1536,0.004688888788223267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,1536,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,1024,0.004016889052258598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,1024,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,3072,256,0.004374222209056218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,768,0.004922666483455234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,768,0.0036311112344264984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,768,0.005711110929648082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,512,0.0033128888656695685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,512,0.00463377767139011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,512,0.005486222190989389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,256,0.004353777815898259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,256,0.002987555538614591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,256,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,128,0.0041644444896115195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,128,0.002801777794957161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,128,0.005153777698675792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,64,0.00263288886182838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,64,0.005127111242877113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,1024,32,0.0026400000270869997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,1024,32,0.005146666533417172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,65536,0.03904799951447381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,65536,0.0665831102265252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,65536,0.06546311246024238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,16384,0.014096889230940076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,16384,0.020900444851981267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,12288,0.016207110550668504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,12288,0.012125333150227865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,12288,0.017287999391555786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,10240,0.01439377831088172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,10240,0.011142222417725457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,10240,0.015729778342776828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,8192,0.012344888514942594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,8192,0.010046222143703038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,8192,0.013307555682129331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,7168,0.010919111470381418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,7168,0.009302222066455418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,7168,0.011864888999197217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,6144,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,6144,0.008555555509196388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,6144,0.0101724440852801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,5120,0.008839111361238692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,5120,0.008544888761308458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,5120,0.009744889206356471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,4096,0.007810666329330868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,4096,0.007083555890454187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,4096,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,3584,0.00740622232357661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,3584,0.007355555891990662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,3584,0.007916444705592262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,3072,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,3072,0.006888000087605582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,3072,0.007547555698288812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,2560,0.006485333459244833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,2560,0.006136888845099344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,2560,0.007128888534175024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,2048,0.006095110956165526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,2048,0.005286222116814719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,2048,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,1536,0.00564177789621883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,1536,0.004658666749795278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,1536,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,1024,0.0039813332259655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,1024,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,10240,0.01779733267095354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,1024,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,768,0.004902222090297275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,768,0.0057546666098965546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,768,0.0036640001667870414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,512,0.0033039999090962936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,512,0.004620444443490771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,512,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,256,0.004300444490379757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,256,0.00295733329322603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,256,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,128,0.0040960001448790235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,128,0.0027866665687825945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,128,0.005164444446563721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,64,0.002647111192345619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,64,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,768,32,0.0026577777332729767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,768,32,0.005164444446563721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,65536,0.028592887851927016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,65536,0.0650764438841078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,16384,0.019935111204783123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,16384,0.011332444846630096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,16384,0.020776000287797716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,12288,0.01590488851070404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,12288,0.009958222508430481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,12288,0.017104888955752056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2560,2560,0.008408000071843466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,10240,0.013644444445768992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,10240,0.009287999735938178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,10240,0.014311111635631986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,8192,0.011535111400816174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,8192,0.008818666968080733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,8192,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,7168,0.010647111468844943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,7168,0.008198222352398766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,7168,0.01091466678513421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,6144,0.007003555695215861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,6144,0.00962577760219574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,6144,0.010105777945783403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,5120,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,5120,0.007099555598364935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,5120,0.009368889033794403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,4096,0.007792889244026608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,4096,0.0063235556913746735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,4096,0.008269333177142674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,3584,0.007370666497283512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,3584,0.007096889118353526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,3584,0.007920000288221572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,3072,0.006519111080302133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,3072,0.007521777517265744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,2560,0.0064533332155810455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,2560,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,2560,0.007134222322040134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,2048,0.006065777606434292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,2048,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,2048,0.006679111056857639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,1536,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,1536,0.00462400002612008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,1536,0.006341333190600078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,1024,0.005106666849719154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,1024,0.003943110919660992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,2048,16384,0.02100533412562476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,1024,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,768,0.0036026665733920205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,768,0.005728000154097875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,512,0.004509333521127701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,512,0.003289777785539627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,512,0.005474666754404704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,256,0.004267555558019214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,256,0.0029315555261241067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,256,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,128,0.004064889003833135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,128,0.002744888886809349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,128,0.0051857779423395796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,64,0.002624888800912433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,64,0.005160888863934411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,512,32,0.0026275554878844153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,512,32,0.005157333281305102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,65536,0.017924444542990792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,65536,0.06565066840913561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,16384,0.01921066641807556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,16384,0.008661333057615492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,65536,0.06479289134343465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,16384,0.01972888906796773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,12288,0.01517955627706316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,12288,0.007532444265153673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,12288,0.01538044379817115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,10240,0.01329066687160068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,10240,0.007098666495747036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,10240,0.01346577786737018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,8192,0.006483555667930179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,8192,0.011445333560307821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,8192,0.01165244479974111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,7168,0.01053244454993142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,7168,0.0063546668324205614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,7168,0.010925333533022137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,6144,0.009553777674833933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,6144,0.006301333506902059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,6144,0.010070222119490305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,5120,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,5120,0.00868000007337994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,5120,0.009284444153308868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,4096,0.007788444558779399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,4096,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,4096,0.008224888808197444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,3584,0.007324444750944774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,3584,0.006747555401590135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,3584,0.007874666816658443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,3072,0.006873777757088344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,3072,0.0064275554484791225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,3072,0.0074933336840735535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,2560,0.006436444405052397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,2560,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,2560,0.007092444433106317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,2048,0.006015111174848344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,2048,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,2048,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,1536,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,1536,0.004609777695602841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,1536,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1536,3584,0.007781333393520779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,1024,0.003993777765168084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,1024,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,768,0.004825777891609403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,768,0.003607999947335985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,768,0.005642666584915585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,512,0.004465777840879229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,512,0.0033031110134389666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,512,0.00544799988468488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,256,0.002971555623743269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,256,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,128,0.004127999974621667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,128,0.0027742222365405825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,128,0.005123555660247803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,64,0.0026337777574857077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,64,0.005144888742102517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,256,32,0.0026133333643277488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,256,32,0.005118222286303838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,65536,0.06606844398710462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,65536,0.013377777404255338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,65536,0.06500889195336236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,16384,0.01901422275437249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,16384,0.006673777682913675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,16384,0.018590221802393597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,12288,0.01518311103185018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,12288,0.00645955569214291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,12288,0.015104000767072042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,10240,0.01332800007528729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,1024,1024,0.0052284445199701525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,10240,0.006671999891599019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,10240,0.013527111046844058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,8192,0.011422221859296164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,8192,0.006516444600290722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,8192,0.011680888632933298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,7168,0.0105253333846728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,7168,0.01092355532778634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,7168,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,6144,0.006096000058783426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,6144,0.009579555855857002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,6144,0.010099555883142684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,5120,0.008692444198661381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,5120,0.006412444429265128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,5120,0.009352888498041365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,4096,0.006039999839332368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,4096,0.008255999949243333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,3584,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,3584,0.006485333459244833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,3584,0.007869333028793335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,3072,0.006823110911581252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,3072,0.006150222072998683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,3072,0.007496000164084964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,2560,0.0063742221229606206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,2560,0.005669333454635408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,2560,0.0070675553547011475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,2048,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,2048,0.005074666606055366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,2048,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,1536,0.005505777895450592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,1536,0.004446222136418025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,1536,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,1024,0.005010666532648934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,768,16384,0.02018311123053233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,1024,0.003839111162556542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,1024,0.0058444444504049085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,768,0.0047457776963710785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,768,0.00564444437623024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,768,0.003532444437344869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,512,0.004425777743260066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,512,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,512,0.005409777578380372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,256,0.0041617775956789655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,256,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,256,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,128,0.004006222304370669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,128,0.0026826666047175727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,128,0.005079999979999331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,64,0.002565333412753211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,64,0.00508711114525795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,128,32,0.0025759999536805684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,128,32,0.005121777868933148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,65536,0.010285333626800114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,16384,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,65536,0.0635991096496582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,16384,0.018566222654448617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,12288,0.015107555521859063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,12288,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,10240,0.006334222025341458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,10240,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,8192,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,8192,0.01163999984661738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,7168,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,7168,0.010967111421955956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,6144,0.006000888844331105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,6144,0.010094222095277574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,5120,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,5120,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,4096,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,4096,0.008247999681366814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,3584,0.006425777657164468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,3584,0.007873777714040544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,3072,0.006145777801672618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,3072,0.007537778053018782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,2560,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,2560,0.007095110913117726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,2048,0.004996444616052839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,2048,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,1536,0.004403555558787452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,1536,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,1024,0.0037537775933742523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,1024,0.0058711109062035876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,768,0.0034337778472238113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,768,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,512,0.003181333343187968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,512,0.005379555539952383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,256,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,256,0.005208000126812193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,128,0.0026586666289303037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,128,0.005084444251325396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,64,0.002514666567246119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,64,0.005103999955786599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,64,32,0.002532444480392668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,64,32,0.005105777747101254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,65536,0.00977244476477305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,65536,0.06232444445292155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,16384,0.006505777852402792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,16384,0.018527110417683918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,12288,0.0063493334584765965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,12288,0.015084443820847405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,10240,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,10240,0.013532444834709167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,8192,0.006203555398517185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,8192,0.011676444775528379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,7168,0.006040888941950268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,7168,0.010915555887752108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,6144,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,6144,0.010064889159467485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,5120,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,5120,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,4096,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,4096,0.008250666989220513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,3584,0.006350222147173352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,3584,0.007871111234029135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,3072,0.006026666611433029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,3072,0.0074782222509384155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,2560,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,2560,0.007088888850477006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,2048,0.005040888984998067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,2048,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,1536,0.004380444271696938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,1536,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,1024,0.0037502224246660867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,1024,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,768,0.003428444473279847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,768,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,512,0.0031386667655573953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,512,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,256,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,256,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,128,0.0026773332307736077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,128,0.00507733349998792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,64,0.0025128889828920364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,64,0.00508355556262864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,16,32,32,0.0025271111064487034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,16,32,32,0.005081777771313985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,65536,0.06650755802790324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,3072,0.0068853331936730286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,16384,0.7123795615302192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,512,768,0.0048142220411035745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,16384,0.3755111164516873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,12288,0.5415289137098525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,16384,0.4041102197435167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,12288,0.28939467006259495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,10240,0.4506977929009332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,12288,0.3049351109398736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,10240,0.23849066098531088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,8192,0.3713946607377794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,8192,0.17831288443671334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,10240,0.24973866674635145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,7168,0.31425510512457955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,8192,0.20621244112650552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,7168,0.18387377262115479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,7168,0.16872444417741564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,6144,0.27040621969434947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,1024,0.005076444397370021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,5120,0.2254577742682563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,6144,0.16761777136060926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,5120,0.12247555785708958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,5120,0.12515644232432047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,4096,0.18366311656104195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,3584,0.15992977884080675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,4096,0.09568799866570367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,3584,0.08512355883916219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,4096,0.110390219423506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,3584,0.088664001888699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,256,256,0.0041786668201287585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,3072,0.13725333743625218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,3072,0.07179288731680976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,2560,0.11546044879489475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,2560,0.06094844473732842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,3072,0.08128266864352755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,2560,0.06493066416846381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,2048,0.0919111106130812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,2048,0.05081599950790405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,2048,0.053039110369152494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,1536,0.04003377755482992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,1536,0.07582222090827094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,1024,0.02857066525353326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,1024,0.047967112726635404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,1536,0.04199377695719401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,768,0.022690667046440974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,768,0.03768800033463372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,1024,0.031481779283947416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,512,0.025891555680169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,768,0.025693333811230127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,512,0.01791466606987847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,512,0.020069332586394418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,128,0.009220444493823582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,256,0.01684533390733931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,256,0.016544888416926067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,128,0.008808000220192803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,64,0.006216889040337668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,65536,32,0.004429333325889376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,128,0.01402311192618476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,64,0.010851555400424533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,65536,32,0.010663111176755695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,16384,0.09407733546362983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,16384,0.17427999443478057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,65536,0.7009555498758951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,12288,0.15708177619510227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,12288,0.0714026689529419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,16384,0.11842400497860378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,10240,0.11001955800586277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,10240,0.06009244256549411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,12288,0.07962577872806124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,65536,0.37342132462395566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,8192,0.08891910976833767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,65536,0.43561066521538633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,8192,0.04951911171277364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,10240,0.07665689124001397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,7168,0.08429244491789077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,7168,0.04437866806983948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,8192,0.05984533495373196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,6144,0.0389048887623681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,7168,0.04963111215167575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,6144,0.07341066996256511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,6144,0.044234666559431285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,5120,0.056932442718082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,5120,0.03522310985459222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,4096,0.046457777420679726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,5120,0.03786311215824551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,4096,0.03104177779621548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,3584,0.026553778184784785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,3584,0.04805688725577461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,3072,0.022428444690174524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,3584,0.028513779242833454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,3072,0.03811288873354594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,3072,0.024826667375034753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,2560,0.030441777573691473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,2560,0.019705777366956074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,2048,0.016720000240537856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,2048,0.02498133314980401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,2560,0.024481778343518574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,2048,0.018910222583346896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,1536,0.01980977753798167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,1536,0.01399466726515028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,1024,0.014000000225173103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,1536,0.01573333309756385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,1024,0.011092444260915121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,1024,0.01275288893116845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,16,128,4096,0.00775733341773351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,768,0.011241777903503843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,768,0.009754666851626502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,768,0.011361777782440186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,512,0.01016533292002148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,256,0.005719999886221356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,256,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,512,0.008447999755541483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,256,0.007122666471534305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,128,0.004407111141416761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,128,0.003363555504216088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,64,0.0030906666070222855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,128,0.006719111154476802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,64,0.006561777657932705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,16384,32,0.003101333354910215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,16384,32,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,16384,0.07206310828526814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,16384,0.13751999537150064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,65536,0.5345039897494847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,16384,0.07772888739903767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,12288,0.06485244300630358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,12288,0.10284444358613755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,65536,0.29213065571255153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,65536,0.3138248920440674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,10240,0.0951404439078437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,10240,0.0471937788857354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,12288,0.05944266584184435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,8192,0.06926933262083265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,10240,0.05093600021468269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,8192,0.040384888648986816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,7168,0.06093421909544203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,8192,0.04186222288343641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,7168,0.034565332863065935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,6144,0.13568888770209417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,6144,0.05217955509821574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,7168,0.037541333172056414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,6144,0.03068355719248454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,6144,0.03769777880774604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,5120,0.04388533367051018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,5120,0.026357332865397137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,4096,0.0358533329433865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,4096,0.022142221530278523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,5120,0.028975112570656672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,4096,0.02385155525472429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,3584,0.03496799866358439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,3584,0.020035554965337116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,3584,0.021785777476098802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,3072,0.017877333694034152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,3072,0.02830577890078227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,3072,0.019849777221679688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,2560,0.02423288921515147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,2560,0.015965332587560017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,2560,0.017470222380426195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,2048,0.019946666227446664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,2048,0.013817778064144982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,2048,0.015103111664454142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,1536,0.015493333339691162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,1536,0.011832888755533429
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,1536,0.012983111043771109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,1024,0.011384000380833944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,768,0.009520000053776635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,1024,0.009619555539555019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,768,0.008591111335489485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,1024,0.01112711098459032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,768,0.009930666950013902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,512,0.007361777954631382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,512,0.006200888918505774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,512,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,256,0.004886222382386525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,256,0.004250666747490565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,256,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,128,0.0033244445092148254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,128,0.006031999985376994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,64,0.003071111109521654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,64,0.006016888966162999
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,12288,32,0.006054222169849608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,12288,32,0.0030826667530669104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,65536,0.45302756627400714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,65536,256,0.013440888788965015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,16384,0.11464177237616645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,16384,0.07170755333370633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,65536,0.22834844059414336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,65536,0.26303288671705455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,12288,0.08621599939134385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,16384,0.06793777810202704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,10240,0.07227289014392428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,12288,0.05184177888764275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,10240,0.04246933261553446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,8192,0.03387199838956197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,10240,0.04504799842834473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,8192,0.058323555522494845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,8192,0.03738311264250014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,7168,0.051671998368369214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,7168,0.03067644437154134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,7168,0.033286223808924355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,6144,0.045306666029824146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,6144,0.02714577813943227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,6144,0.029459555943806965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,5120,0.03796089026663039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,5120,0.02386577758524153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,5120,0.025913778278562758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,4096,0.030869334936141968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,4096,0.01979200045267741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,3584,0.02740977704524994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,4096,0.021405332618289526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,3584,0.01831733352608151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,3584,0.019662222928471036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,3072,0.0240657776594162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,3072,0.016540444559521146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,3072,0.017816889617178176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,2560,0.020629333125220407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,2560,0.014755555325084262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,2560,0.015788444214397006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,2048,0.016851555969980028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,2048,0.013954665925767688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,2048,0.012703999876976013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,1536,0.013415999710559845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,1536,0.011015111373530494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,1536,0.012160888976520963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,1024,0.009862222605281407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,1024,0.009212444225947062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,1024,0.0105608892109659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,768,0.008220444122950235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,768,0.008247999681366814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,768,0.009744000103738572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,512,0.0064897777305708984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,512,0.0054560001525614
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,512,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,256,0.0034728890491856467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,256,0.004956444518433677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,256,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,128,0.00309599998096625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,128,0.004480888860093223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,128,0.006047111004590988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,64,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,10240,32,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,64,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,10240,32,0.005986666513813867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,65536,0.353603548473782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,4096,0.027811557054519653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,16384,0.05707199705971611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,65536,0.18648266792297363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,16384,0.08798399898740981
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,65536,0.23001421822441948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,12288,0.06757511032952203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,16384,0.061088886525895864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,12288,0.039975110027525164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,16384,512,0.008231110870838165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,12288,0.04722133278846741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,10240,0.05649155378341675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,10240,0.040159109565946795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,8192,0.046113779147466026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,7168,0.04080977704789903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,7168,0.025204444924990337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,8192,0.03261155552334256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,6144,0.035940445131725736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,7168,0.030772444274690416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,6144,0.022777777579095628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,5120,0.030365334616767034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,5120,0.01978577839003669
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,6144,0.027297778262032404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,4096,0.0248391115003162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,5120,0.02312266661061181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,4096,0.016861332787407767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,3584,0.022198221749729578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,4096,0.019966221517986722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,3584,0.015442666080262927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,3072,0.019545776976479422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,3584,0.017887110511461895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,3072,0.014124444789356656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,3072,0.016677333248986136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,2560,0.01679733395576477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,2560,0.012692444854312472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,2048,0.013897778259383308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,2560,0.014914666612943014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,2048,0.01111822244193819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,1536,0.01124888906876246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,2048,0.013245333400037555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,1536,0.011602666642930774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,1024,0.008200888832410177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,1024,0.008449777960777283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,1024,0.010104888843165504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,768,0.006525333142942852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,768,0.007103111180994246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,768,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,512,0.005720888988839255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,512,0.005048888838953442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,512,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,256,0.004722666823201709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,256,0.0063564446237352155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,256,0.0033893332713180115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,128,0.00309244439833694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,128,0.004310222135649787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,128,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,64,0.0030782222747802734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,64,0.005872000008821487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,8192,32,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,8192,32,0.006001777946949005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,65536,0.30613332324557835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,16384,0.05284622311592102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,65536,0.18006932735443115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,65536,0.18566577964358857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,16384,0.08119822210735746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,12288,0.0632186664475335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,12288,0.037181334363089666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,16384,0.04987733231650459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,10240,0.055296887954076133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,10240,0.0320355561044481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,12288,0.03828710979885525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,8192,0.04402755697568258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,10240,0.033490667740503945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,8192,0.026161778304311965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,8192,0.027709333433045283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,7168,0.0236835562520557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,7168,0.03941866755485535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,6144,0.03514488869243198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,7168,0.025037333369255066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,6144,0.021193777521451313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,6144,0.02239022155602773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,5120,0.029861334297392104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,5120,0.018349332941903006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,5120,0.020046222541067336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,4096,0.015654222832785714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,4096,0.022484444909625586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,4096,0.016565332810084026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,3584,0.014497778481907315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,3584,0.02010844482315911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,3584,0.015455999308162265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,3072,0.013298666311634911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,3072,0.01790577835506863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,3072,0.014275555809338888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,2560,0.011901333100265928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,2560,0.015303111738628812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,2560,0.01293333371480306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,12288,128,0.004376000000370873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,2048,0.010410666465759277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,2048,0.012836444709036084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,2048,0.011505777637163797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,1536,0.010438222024175854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,1536,0.009272000028027428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,1536,0.010297777752081553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,1024,0.007914666500356462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,1024,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,1024,0.008824889030721452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,768,0.006684444430801604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,768,0.005476444545719359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,768,0.006766222417354584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,512,0.005209777918126848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,512,0.005108444641033809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,512,0.006164444403515921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,256,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,256,0.0046693334976832075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,256,0.005713777823580637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,128,0.002967111145456632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,7168,128,0.004212444441186057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,128,0.005480000128348668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,64,0.002777777819169892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,64,0.00535733335547977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,7168,32,0.0028657778683635923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,7168,32,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,65536,0.26552266544765896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,16384,0.047078222036361694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,65536,0.15005866686503092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,65536,0.15910666518741184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,16384,0.0734986662864685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,16384,0.04635644290182325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,12288,0.05594311157862345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,12288,0.03264977865748935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,10240,12288,0.049676444795396596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,10240,0.045365333557128906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,12288,0.03363377849260966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,10240,0.02805688977241516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,10240,0.03202222122086419
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,8192,0.037598222494125366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,8192,0.025008888708220586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,8192,0.02312799957063463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,7168,0.03426488902833726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,7168,0.02330400049686432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,6144,0.029000888268152874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,6144,0.019104889697498746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,5120,0.016751999656359356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,6144,0.02368533280160692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,5120,0.02460533380508423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,4096,0.022600889205932617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,5120,0.018368888232443068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,4096,0.014398222168286642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,4096,0.015003555350833468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,3584,0.01862133377128177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,3584,0.015581332974963717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,3584,0.014184888866212634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,3072,0.016387555334303115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,3072,0.01237511138121287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,2560,0.014117333624098035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,3072,0.015885333220163982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,2560,0.011192888849311404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,2560,0.01220533334546619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,2048,0.009928888744778102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,2048,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,2048,0.010830221904648675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,1536,0.009692444569534725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,1536,0.012396444049146442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,1536,0.00886399961180157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,1024,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,1024,0.007382222347789341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,768,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,1024,0.010611555642551847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,768,0.006481777876615524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,768,0.005445333404673471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,512,0.00573333352804184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,512,0.006033777776691649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,512,0.005072000126043956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,256,0.004776000148720211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,256,0.004499555461936527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,256,0.005647999958859549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,128,0.002970666728085942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,128,0.004386666748258802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,64,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,128,0.007712889048788283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,64,0.005366222312053044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,6144,32,0.002768888862596618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,6144,32,0.005382222019963794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,65536,0.2259297900729709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,16384,0.04367466767628988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,65536,0.1271911064783732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,65536,0.16233689255184597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,16384,0.057824002371893994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,12288,0.04442666636572944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,16384,0.036912888288497925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,12288,0.041464000940322876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,12288,0.029472000069088403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,10240,0.026224889689021643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,10240,0.04565599891874525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,10240,0.02566755480236477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,8192,0.03693777653906081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,8192,0.022146667043368023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,8192,0.021497777766651575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,7168,0.027265777190526325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,7168,0.02835644284884135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,7168,0.019646222392718
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,6144,0.017860444055663217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,6144,0.028981334633297388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,5120,0.02494577732351091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,6144,0.017832888497246634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,5120,0.01591199967596266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,4096,0.01705333259370592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,5120,0.02221600049071842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,4096,0.013555555707878537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,4096,0.01369866645998425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,3584,0.01538222200340695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,3584,0.01733955575360192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,3584,0.01292622254954444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,3072,0.016689777374267578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,10240,0.03411022159788344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,3072,0.011517333487669626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,3072,0.01198755535814497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,2560,0.011864888999197217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,2560,0.011025778121418424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,2048,0.010089777410030365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,2048,0.009387555221716562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,2048,0.010048888623714447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,8192,0.028231110837724473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,1536,0.00849955528974533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,1536,0.008452444440788692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,1536,0.009017777939637503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,1024,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,1024,0.006127111199829314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,1024,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,768,0.005744888964626525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,768,0.005381333331267039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,768,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,512,0.004973333328962326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,512,0.003456888927353753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,512,0.0058648888435628675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,256,0.00309599998096625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,256,0.005548444473081165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,128,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,128,0.0041955556306574075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,64,0.0027280000762807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,64,0.0052595556610160405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,128,0.005373333477311664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,5120,32,0.002716444432735443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,5120,32,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,65536,0.18431733714209664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,65536,0.09824444188012017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,16384,0.03276177909639146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,65536,0.10438755485746597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,8192,1536,0.00963288876745436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,16384,0.04866755670971341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,16384,0.031201776531007554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,12288,0.0384897788365682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,12288,0.02548711167441474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,12288,0.024120888776249353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,10240,0.02124711043304867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,10240,0.03271911210483975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,8192,0.027148443791601393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,10240,0.022302221920755174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,8192,0.01867644488811493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,7168,0.02458400030930837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,7168,0.016413332687483895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,7168,0.017263111140992906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,6144,0.021945777866575453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,6144,0.014986667368147107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,6144,0.01579466627703773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,5120,0.019275556008021038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,5120,0.013405332962671915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,5120,0.01440355512830946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,4096,0.014217777384652032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,4096,0.011585777832402123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,4096,0.012314666476514606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,3584,0.012763555679056378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,3584,0.011603555745548673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,3584,0.010957332948843638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,3072,0.011529777612951068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,3072,0.010858666565683154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,2560,0.010064889159467485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,2560,0.009380444056457942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,2560,0.010191111101044549
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,2048,0.008454221818182204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,2048,0.00868088917599784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,2048,0.009321778184837764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,1536,0.007391110890441471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,1536,0.0069146665434042616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,1536,0.007737778127193451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,1024,0.005457777943876054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,1024,0.006001777946949005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,1024,0.006143110907740063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,768,0.005154666801293691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,768,0.004855111241340637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,768,0.005859555469618903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,512,0.004957333207130432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,512,0.0033884443756606844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,512,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,256,0.004567111117972268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,256,0.0030746666921509635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,256,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,128,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,128,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,128,0.005220444252093633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,64,0.0026373333401150177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,64,0.005222222043408289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,4096,32,0.0026417778184016547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,4096,32,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,65536,0.1565964486863878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,65536,0.09457955757776897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,16384,0.028639998700883653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,16384,0.0435733331574334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,65536,0.104184885819753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,16384,0.028728001647525366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,12288,0.023054222265879314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,12288,0.03395911057790121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,12288,0.0237128883600235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,10240,0.020399111840460036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,10240,0.029502222935358684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,10240,0.02069511181778378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,8192,0.017176000608338248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,8192,0.024586665961477492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,8192,0.017249777913093567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,7168,0.015643555257055495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,7168,0.022395556171735127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,7168,0.01621955633163452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,6144,0.014205333259370593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,6144,0.019890666007995605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,6144,0.014926221635606555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,5120,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,5120,0.012822222378518848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,5120,0.013632000320487551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,4096,0.010947555303573608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,4096,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,4096,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,3584,0.010432889064153036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,3584,0.011763555308183035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,3584,0.011121778024567498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,3072,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,3072,0.01054044481780794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,3072,0.010478222535716163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,2560,0.009116444322797988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,2560,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,2560,0.00980533328321245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,2048,0.00816888858874639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,2048,0.008236444658703273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,2048,0.0090506664580769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,1536,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,1536,0.006032889087994893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,1536,0.007378666765160031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,1024,0.005366222312053044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,1024,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,1024,0.006027555714050929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,768,0.005180444568395615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,768,0.004013333469629288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,768,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,512,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,512,0.0034213333080212274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,512,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,256,0.003039999968475766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,256,0.004553777890072929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,256,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,128,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3584,128,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,128,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,64,0.0026577777332729767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,64,0.005186666631036335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3584,32,0.0026337777574857077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3584,32,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,65536,0.1389626661936442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,65536,0.07717333237330119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,65536,0.08656977944903904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,16384,0.028839998775058325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,16384,0.038895110289255776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,12288,0.022112000319692824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,12288,0.03049066662788391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,16384,0.026498667067951624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,10240,0.026610665851169165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,12288,0.021692444880803425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,10240,0.01959911154376136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,8192,0.022100443641344707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,10240,0.019346666004922655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,8192,0.017115554875797696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,7168,0.020044444335831534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,8192,0.0162462227874332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,7168,0.015288000305493673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,6144,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,6144,0.01775733298725552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,6144,0.014090667168299357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,6144,7168,0.021034666233592566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,5120,0.012353777885437012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,5120,0.01292622254954444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,5120,0.01571733256181081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,4096,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,4096,0.010767111347781287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,4096,0.011228444675604502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,3584,0.010077333284748925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,3584,0.013001778059535556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,3584,0.010709332923094431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,3072,0.011733333269755045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,3072,0.010031110710567897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,2560,0.008832000195980072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,2560,0.008807111117574904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,2560,0.009385777844323052
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,2048,0.007173333730962541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,2048,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,2048,0.007796444826655918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,1536,0.005800888770156437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,1536,0.006575111299753189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,1536,0.006493333313200209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,1024,0.0053546664615472155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,1024,0.005107555538415909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,1024,0.006043555421961679
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,768,0.00508622204264005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,768,0.003727999826272329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,768,0.005867555737495422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,512,0.0047591109242704184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,512,0.003387555480003357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,512,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,256,0.004382222063011593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,256,0.0030586665703190696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,256,0.005411555369695027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,128,0.004096888833575779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,128,0.0027795556104845474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,128,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,64,0.0026293332791990707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,64,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,3072,32,0.0026186667382717133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,3072,32,0.005240889059172736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,65536,0.07595377498202853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,65536,0.11569778124491374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,65536,0.08043999804390801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,2560,0.010536889235178629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,16384,0.03286666671435038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,16384,0.02381333377626207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,5120,256,0.0046773333516385825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,12288,0.026300445199012756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,16384,0.024632889363500807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,12288,0.02017866737312741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,12288,0.020344000723626878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,10240,0.017921777235137094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,10240,0.02287111183007558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,10240,0.018062222335073683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,8192,0.019263111882739596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,8192,0.015575110912322998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,8192,0.01795377830664317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,7168,0.01752355529202355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,7168,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,6144,0.015555555621782938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,7168,0.01443377799457974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,6144,0.013433777623706393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,5120,0.011528888510333167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,5120,0.013944000005722046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,5120,0.012366222010718452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,4096,0.010248888697889116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,4096,0.012085333466529846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,4096,0.010761777559916178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,3584,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,3584,0.009748444788985783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,3584,0.01036711119943195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,3072,0.010497777826256223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,3072,0.009229333036475712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,3072,0.009698666632175446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,2560,0.007846222155623967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,2560,0.00811111099190182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,2560,0.008716444174448649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,2048,0.007060444189442529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,2048,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,2048,0.0068897778789202375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,1536,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,1536,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,1024,0.0042364444169733255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,1024,0.005643555687533484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,1024,0.006063999815119638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,768,0.005052444421582752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,768,0.003698666476541095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,768,0.0058151111006736755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,512,0.0047120000753137796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,512,0.0033440000067154565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,512,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,256,0.003031999907559819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,256,0.004338666796684265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,256,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,128,0.004025777594910728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,128,0.002817777709828483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,64,0.0026693333768182327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,128,0.005163555757866966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,64,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,8192,0.017842666970358956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2560,32,0.0026613333159022858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2560,32,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,65536,0.09393333064185248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,4096,3072,0.010152888794740042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,65536,0.06812533405092028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,16384,0.021055999729368422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,16384,0.0270106659995185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,12288,0.017138666576809354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,16384,0.02276977731121911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,65536,0.07278399997287326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,12288,0.021869333253966436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,12288,0.018929777873886954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,10240,0.019267555740144517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,10240,0.0169804443915685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,8192,0.013302221894264221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,8192,0.01592177814907498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,8192,0.01460533340771993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,7168,0.01240533341964086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,7168,0.014645333091417948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,7168,0.013647110925780402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,6144,0.011517333487669626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,6144,0.013230221966902414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,6144,0.012668444050682915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,5120,0.01053600013256073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,5120,0.012196444802814059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,5120,0.011724444727102915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,4096,0.009553777674833933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,4096,0.010664888554149203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,4096,0.010207110808955299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,3584,0.008813333180215623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,3584,0.010091555615266165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,3584,0.009749333063761393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,3072,0.007687999970383114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,3072,0.00944977750380834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,3072,0.008608889248636033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,2560,0.006686222222116258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,2560,0.007443555527263218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,2560,0.007232888705200619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,2048,0.006207110981146495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,2048,0.006666666517655055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,2048,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,1536,0.005724444571468566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,1536,0.005294222384691238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,1536,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,1024,0.005202666752868229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,1024,0.003940444439649582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,1024,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,768,0.005045333256324132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,768,0.003656889001528422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,768,0.005680888891220093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,512,0.0047253333032131195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,512,0.0032960000551409195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,512,0.005467555589146084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,256,0.004395555704832077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,256,0.00300177786913183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,256,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,128,0.0041759999261962045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,128,0.0027502222607533136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,128,0.005139555368158552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,64,0.0026071110947264563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,64,0.005135111096832487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,2048,32,0.002601777720782492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,2048,32,0.005123555660247803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,65536,0.07063821951548259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,65536,0.06720444228914049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,16384,0.020768889122539096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,65536,0.06813510921266344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,16384,0.022152000003390845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,16384,0.021537777450349595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,12288,0.016691555579503376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,12288,0.018041777941915724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,12288,0.017973333597183228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,10240,0.014814222852389017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,10240,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,10240,0.016252444850073922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,8192,0.012866666747464074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,8192,0.013584000368913015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,8192,0.013924444715181986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,7168,0.011946666571829053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,7168,0.012700444294346703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,7168,0.01314311143424776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,6144,0.01165066659450531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,6144,0.011055111057228513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,6144,0.012196444802814059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,5120,0.010152888794740042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,5120,0.010695111420419482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,5120,0.011385777758227454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,4096,0.00866844422287411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,4096,0.009512888888518015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,4096,0.009357333183288574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,3584,0.009107555780145857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,3584,0.007733333441946242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,3584,0.008473777936564552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,3072,0.0070497774415545994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,3072,0.008589333130253686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,3072,0.00756177802880605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,2560,0.006525333142942852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,2560,0.007088888850477006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,2560,0.007116444408893585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,2048,0.006076444354322221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,2048,0.006170666466156642
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,2048,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,1536,0.004615111069546806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,1536,0.005678222411208683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,1536,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,1024,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,1024,0.00397511116332478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,1024,0.0058666666348775225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,768,0.004874666531880696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,768,0.0036106668412685394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,768,0.005714666512277391
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,512,0.004668444395065308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,512,0.003289777785539627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,512,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,256,0.002983111060327954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,256,0.004362666772471534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,256,0.005304000029961268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1536,128,0.004050666673315896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,128,0.0027582223216692605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,128,0.005147555636035072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,64,0.002577777744995223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,64,0.005166222237878376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1536,32,0.0026142222599850762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1536,32,0.00517155561182234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,65536,0.049505776829189725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,65536,0.06575822167926364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,16384,0.02017244365480211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,65536,0.06556089056862725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,16384,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,16384,0.02087644404835171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,12288,0.0163137780295478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,12288,0.013759111364682516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,12288,0.017398221625222098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,10240,0.012462221913867526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,10240,0.014412444498803882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,10240,0.01573422220018175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,8192,0.012503999802801343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,8192,0.010899555351999072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,8192,0.013423110875818463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,7168,0.010294222169452244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,7168,0.011648000114493899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,7168,0.012670222255918713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,6144,0.010464888479974536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,6144,0.0092604441775216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,6144,0.011077333655622272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,5120,0.00887111077706019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,5120,0.008912888665994009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,5120,0.009673777553770278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,4096,0.00835111074977451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,4096,0.007833778030342525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,4096,0.008432889150248634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,3584,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,3584,0.00740533322095871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,3584,0.007903999752468532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,3072,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,3072,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,3072,0.007500444849332173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,2560,0.006455111006895701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,2560,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,2560,0.007107555866241455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,7168,0.015421332584487068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,2048,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,2048,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,1536,0.004631999880075455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,2048,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,1536,0.005642666584915585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,1536,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,1024,0.003929777691761653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,1024,0.005151111218664381
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,1024,0.005802666561471091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,768,0.0035742223262786865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,768,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,768,0.004895110925038655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,512,0.003248888792263137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,512,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,512,0.0054551110499435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,256,0.0029484445436133277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,256,0.004290666845109728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,256,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,128,0.002732444347606765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,1024,128,0.004044444610675176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,128,0.005141333159473208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,64,0.0026133333643277488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,64,0.005136888888147143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,1024,32,0.0026195556339290407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,1024,32,0.005123555660247803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,3072,3072,0.009465778039561378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,65536,0.038253333833482527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,65536,0.06545955604977079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,16384,0.013937777943081327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,65536,0.06519466638565063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,12288,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,12288,0.011796444654464722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,16384,0.02079288827048408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,12288,0.01721155477894677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,10240,0.01091822236776352
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,10240,0.014324444863531323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,8192,0.009881777895821465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,10240,0.015572445260153877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,8192,0.011971555650234222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,8192,0.013056000073750814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,7168,0.010913777682516309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,7168,0.009166222479608323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,7168,0.011518222590287527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,6144,0.008440000315507254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,6144,0.00979022185007731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,6144,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,5120,0.008700444466537898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,5120,0.008174222376611497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,5120,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,4096,0.007810666329330868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,4096,0.006887110984987683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,4096,0.00831733312871721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,3584,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,3584,0.007377777662542131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,3584,0.007920888562997183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,3072,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,3072,0.0069217777086628815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,3072,0.0075457774930530125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,2560,0.005990222096443176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,2560,0.0064560001095136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,2560,0.007173333730962541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,2048,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,2048,0.006119110931952794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,2048,0.006671111202902264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,1536,0.0046364445653226645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,1536,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,1536,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,1024,0.003936000168323517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,1024,0.005162666655249066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,1024,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,768,0.003592888928121991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,768,0.004880889008442561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,768,0.005715555614895291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,512,0.0032835555159383347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,512,0.005518222020732031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,256,0.0029279999434947968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,256,0.004279110994603899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,256,0.005312888986534542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,128,0.002699555622206794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,128,0.003957333250178231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,128,0.005414222263627582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,64,0.002597333242495855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,64,0.005183111048407025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,768,32,0.0026195556339290407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,768,32,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,65536,0.027954667806625366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,65536,0.06534400251176622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,16384,0.019902222686343722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,65536,0.06485511196984185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,16384,0.011032000184059143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,16384,0.020735111501481798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,12288,0.01587377819750044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,12288,0.009744000103738572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,12288,0.016902221573723685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,10240,0.009157333109113906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,10240,0.013647110925780402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,10240,0.013813333378897773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,8192,0.011476444701353708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,8192,0.008584000170230865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,8192,0.01183377785815133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,7168,0.010494222243626913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,7168,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,7168,0.010888000329335531
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,6144,0.009610666996902889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,6144,0.007063999772071838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,6144,0.01015466699997584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,5120,0.008647110727098253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,5120,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,5120,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,4096,0.00776533368561003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,4096,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,4096,0.008296888735559251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,3584,0.006807111203670502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,3584,0.007879111501905654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,3072,0.006872889068391588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,3072,0.006426666759782367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,3072,0.007520889242490132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,2560,0.006413333531883027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,2560,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,2560,0.007113777928882175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,6144,0.012979555461141797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,2048,0.005305777821275923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,2048,0.006664888726340399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,1536,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,1536,0.004611555486917496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,1024,0.005118222286303838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,1024,0.003950222084919612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,1024,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2560,1536,0.0058168888919883305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,768,0.004830222162935468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,768,0.0035742223262786865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,768,0.005671111245950063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,512,0.003289777785539627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,512,0.004493333399295807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,512,0.005486222190989389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,256,0.004221333397759331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,256,0.0029404444826973807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,256,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,128,0.003915555361244413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,128,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,128,0.005136000199450387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,64,0.0026142222599850762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,64,0.005176888985766305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,512,32,0.0026364444444576898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,512,32,0.005116444494989184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,65536,0.017364444004164804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,65536,0.06609600120120578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,16384,0.01924800044960446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,65536,0.06455555227067736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,16384,0.00851111114025116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,16384,0.019348444210158456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,12288,0.015240000353919135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,12288,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,12288,0.015138667490747241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,10240,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,10240,0.013639999760521783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,8192,0.011543110840850405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,8192,0.0064399999876817065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,8192,0.011714666253990598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,7168,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,7168,0.006360888895061281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,7168,0.010944888823562197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,6144,0.009661333428488838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,6144,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,6144,0.01014311114947001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,2048,10240,0.01531733406914605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,5120,0.008766222331258986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,5120,0.009383999639087254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,5120,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,4096,0.006207110981146495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,4096,0.007814221911960179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,4096,0.008207110895050896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,3584,0.007400000260935889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,3584,0.006688000013430913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,3584,0.007888000044557784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,3072,0.006370666540331311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,3072,0.007528000407748752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,2560,0.0063902222447925145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,2560,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,2560,0.00700800038046307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,2048,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,2048,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,2048,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,1536,0.005531555662552516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,1536,0.004592888885074192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,1536,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,1024,0.005025777965784073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,1024,0.003918222255176968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,1024,0.005762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,768,0.004790222065316306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,768,0.0035911111368073356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,768,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,512,0.004499555461936527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,512,0.0032542221662071017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,512,0.005379555539952383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,256,0.0041475556790828705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,256,0.002895111011134254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,256,0.00517866677708096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,128,0.003919110943873723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,128,0.002672000063790215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,128,0.005056000004212062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,64,0.002581333327624533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,64,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,256,32,0.0026071110947264563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,256,32,0.005059555586841372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,65536,0.06502044200897217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,65536,0.012352888782819113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,65536,0.06432888905207317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,16384,0.018711111611790128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,16384,0.006414222220579783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,16384,0.01828177769978841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,12288,0.01497244503762987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,12288,0.006178666734033161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,12288,0.014927110738224454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,10240,0.01314399970902337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,10240,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,10240,0.013337777720557319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,8192,0.011348444554540845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,8192,0.0064088888466358185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,8192,0.011537777880827585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,7168,0.006159111029571957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,7168,0.010759111079904767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,6144,0.009549332989586724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,6144,0.00599911105301645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,6144,0.009974222216341231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,5120,0.008591999610265097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,5120,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,5120,0.009222221871217092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,4096,0.007688889073001013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,4096,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,4096,0.008170666793982187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,3584,0.007229333122571309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,3584,0.0063768890168931745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,3584,0.007787555456161499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,3072,0.006783999916579988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,3072,0.006048000107208888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,3072,0.0074782222509384155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,2560,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,2560,0.005702222386995952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,2560,0.007114667031500075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,2048,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,2048,0.005074666606055366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,2048,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,1536,0.005501333210203383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,1536,0.004466666529575984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,1536,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,1024,0.005003555367390315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,1024,0.0037911112109820047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,1024,0.0058133333093590206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,768,0.004724444614516364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,768,0.003489777859714296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,768,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,512,0.004422222160630756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,512,0.0031724443866146933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,512,0.005395555661784277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,256,0.0041253334946102565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,256,0.002862222285734283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,256,0.005236444373925527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,128,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,128,0.0027191111197074256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,128,0.005101333475775189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,64,0.0025884444928831523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,64,0.005124444348944558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,128,32,0.002575111058023241
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,128,32,0.005123555660247803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,65536,0.009791110952695211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,65536,0.06287022431691487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,16384,0.006130666782458623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,16384,0.018519111805491976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,12288,0.00601333338353369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,16384,0.020012444920010038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,12288,0.014891554911931356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,10240,0.006515555497672822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,10240,0.013340444200568728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,8192,0.006167999986145232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,7168,0.006046222315894232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,8192,0.011508444117175208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,6144,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,7168,0.010777778095669217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,6144,0.009963555468453301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,5120,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,4096,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,5120,0.009222221871217092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,3584,0.006344888773229387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,4096,0.008197333249780867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,3584,0.007744889292452071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,3072,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,3072,0.007376888559924231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,2560,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,2560,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,2048,0.004930666751331753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,2048,0.006517333288987477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,1536,0.004337777694066365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,1536,0.006133333262470033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,1024,0.0037368887828456033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,1024,0.0057653333577844836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,768,0.0034515555534097883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,768,512,0.004578666554556953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,512,0.0031644445326593188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,768,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,512,0.005356444252861871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,256,0.002862222285734283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,256,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,128,0.005103999955786599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,128,0.0026062221990691293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,64,0.0025066667132907445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,64,0.00507377791735861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,64,32,0.002479111154874166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,64,32,0.005049777941571342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,65536,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,16384,0.006089777582221561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,16384,0.01830844415558709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,12288,0.006083555519580841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,65536,0.059952888223859996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,10240,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,12288,0.014925332532988654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,10240,0.013405332962671915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,8192,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,8192,0.011619555453459421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,7168,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,6144,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,7168,0.010833777487277985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,6144,0.009949333137936061
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,5120,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,5120,0.00923644420173433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,4096,0.005836444596449534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,4096,0.008180444439252218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,3584,0.00777422222826216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,3584,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,3072,0.00601244428091579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,3072,0.00741511086622874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,2560,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,2560,0.0070159998204973005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,2048,0.004906666775544484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,2048,0.006566222343179915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,1536,0.004374222209056218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,1536,0.006124444305896759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,1024,0.003727999826272329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,1024,0.005731555736727185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,768,0.0034435554924938413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,768,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,512,0.003128000017669466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,512,0.0053431110249625305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,256,0.0028151110228565005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,256,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,128,0.00263288886182838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,128,0.005069333232111401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,64,0.0025048889219760895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,64,0.005029333134492238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,8,32,32,0.002479111154874166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,8,32,32,0.005050666630268097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,3584,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,16384,0.7126755714416504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,512,2048,0.006053333481152852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,16384,0.3757199976179335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,12288,0.5417973200480143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,12288,0.3141520023345947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,16384,0.39855111969841855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,10240,0.45009602440728086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,12288,0.3077190981970893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,10240,0.25940267244974774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,10240,0.013218666944238873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,8192,0.3607333236270481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,10240,0.26769423484802246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,8192,0.2083439959420098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,7168,0.31520533561706543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,256,3072,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,8192,0.22171289390987822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,7168,0.17923377619849312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,6144,0.27107376522488064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,6144,0.13533243868086073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,7168,0.1950604385799832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,5120,0.22458932134840223
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,4096,0.1829813321431478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,6144,0.1881635586420695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,5120,0.13226400481330022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,4096,0.1073955562379625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,4096,0.09806844260957505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,3584,0.15908888975779215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,3584,0.09021333191129897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,3072,0.13704355557759604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,2560,0.11415910720825195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,3072,0.07120444377263387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,3584,0.10620266861385769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,2560,0.06088888645172119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,3072,0.07529689206017388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,2048,0.09094666772418553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,2560,0.06470844480726454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,2048,0.0498542222711775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,1536,0.07663289043638441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,1536,0.04500888784726461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,1024,0.028191109498341877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,2048,0.06266311142179701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,1024,0.04756888747215271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,1536,0.04420711265669929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,768,0.03735466798146566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,768,0.022409778502252366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,1024,0.03048266636000739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,512,0.017380444539917838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,512,0.02570488883389367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,256,0.012240000069141388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,256,0.01641155613793267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,768,0.030766222212049697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,512,0.020015999674797058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,256,0.016680000556839835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,128,0.008827555510732863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,128,0.009115555220180089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,128,0.013975110318925647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,64,0.010673777924643623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,64,0.010105777945783403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,65536,32,0.010588444769382477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,65536,32,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,16384,0.17427466975318062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,16384,0.10248977608150905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,16384,0.10769689083099365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,12288,0.07129422161314222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,12288,0.1349440018335978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,12288,0.0809217757648892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,8,128,7168,0.010586666564146677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,65536,0.7000862227545844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,10240,0.06484266784456041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,10240,0.11196711328294541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,8192,0.08913866678873698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,8192,0.056008888615502246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,7168,0.07829778061972724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,10240,0.06827733251783583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,7168,0.0528417792585161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,8192,0.056503110461764865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,6144,0.06843111250135633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,6144,0.0387733346886105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,7168,0.050138665570153125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,6144,0.04938133226500618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,5120,0.05671733617782593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,4096,0.04735733403099907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,5120,0.03767377800411648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,65536,0.37413777245415586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,4096,0.027543110979927912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,4096,0.030964444080988567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,3584,0.04093955622778998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,65536,0.43633421262105304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,3072,0.03573866685231527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,3072,0.022257778379652236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,3584,0.028174221515655518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,2560,0.030226667722066242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,2560,0.019595555133289762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,2560,0.0217875556813346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,3072,0.026120889517996047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,2048,0.02474311159716712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,2048,0.01661155621210734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,2048,0.018681777848137748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,1536,0.01958933307064904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,1536,0.014362666341993542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,1024,0.013911111487282647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,1536,0.015697777271270752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,1024,0.010951999988820819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,768,0.011172444456153445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,1024,0.013223110801643796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,768,0.011324444578753578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,512,0.00813422186507119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,512,0.009936889012654623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,512,0.008914666871229807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,256,0.004980444494220945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,256,0.005685333162546158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,128,0.003363555504216088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,256,0.007191110816266801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,128,0.004412444515360726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,128,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,64,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,32,0.003048000029391713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,16384,64,0.003031999907559819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,16384,32,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,65536,0.5337013138665093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,65536,0.31503645579020184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,16384,0.1370560063256158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,16384,0.07752533091439141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,12288,0.055456890000237354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,12288,0.10239466693666245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,12288,0.05986844168768989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,10240,0.04686844348907471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,10240,0.08333333333333333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,10240,0.05076977941724989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,8192,0.038667556312349104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,8192,0.06764800018734403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,8192,0.04157688882615831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,7168,0.03425511055522495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,65536,5120,0.11340888341267903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,7168,0.06070044305589464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,6144,0.051995555559794106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,7168,0.03724977705213759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,6144,0.030356443590588037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,5120,0.044681777556737266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,5120,0.02602844436963399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,6144,0.0329733325375451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,4096,0.03614577651023865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,5120,0.028592000404993694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,4096,0.021899556120236714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,3584,0.020068445139461093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,4096,0.023798222343126934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,3584,0.03205777870284186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,3072,0.028212444649802312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,3584,0.021659556362364028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,3072,0.017802667286660936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,2560,0.023895111348893907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,3072,0.019513777560657926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,2560,0.017126222451527912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,2048,0.013667555318938361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,2048,0.01964000033007728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,2048,0.014933332800865173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,1536,0.011582222249772815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,1536,0.015583111180199517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,1536,0.012865777644846173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,1024,0.009509333305888707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,1024,0.011382222175598145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,1024,0.010964444114102257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,768,0.008470222353935242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,768,0.009534222384293875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,768,0.009939555492666032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,512,0.006435555716355641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,512,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,512,0.008030222521887885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,256,0.0047973332305749255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,256,0.0041599998043643106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,256,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,128,0.004348444441954295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,128,0.003334222154484855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,64,0.003080888961752256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,128,0.006031111296680238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,64,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,12288,32,0.0030951110853089225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,12288,32,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,65536,0.43965866830613876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,65536,0.2442399925655789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,16384,0.06707733207278781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,5120,0.03316444489690993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,3584,0.025137777129809063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,16384,0.11258310741848415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,16384,0.06807021962271796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,65536,0.29496977064344615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,12288,0.04847200049294365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,16384,768,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,12288,0.08611733383602566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,10240,0.043791999419530235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,10240,0.0720720026228163
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,12288,0.051959110630883105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,10240,0.04459911253717211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,8192,0.05804710918002658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,8192,0.036143110858069524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,7168,0.05121066504054599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,8192,0.03693600164519416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,7168,0.03038933210902744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,6144,0.04496533340877957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,7168,0.03710222244262695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,6144,0.029122667180167303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,5120,0.02344266739156511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,5120,0.03962666789690653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,5120,0.025606221622890894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,4096,0.019543111324310303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,4096,0.03204977843496535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,4096,0.021393777595625982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,3584,0.02878311276435852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,3584,0.018014222383499146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,3072,0.023892444041040208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,3584,0.022192888789706763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,3072,0.01761155492729611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,2560,0.02032444377740224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,2560,0.015657777587572735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,2048,0.01255733354224099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,2048,0.016764443781640794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,2048,0.013883555928866068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,1536,0.011725333001878528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,1536,0.013272888958454132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,1536,0.012068444656001197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,1024,0.009089777866999308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,1024,0.009864889085292816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,1024,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,768,0.008184000021881526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,768,0.008216888540320927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,768,0.009523555636405945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,512,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,512,0.006434666613737742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,512,0.007038222418891058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,256,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,256,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,256,0.0035253332720862497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,128,0.003136000078585413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,128,0.004399999976158142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,64,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,128,0.006024888820118374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,64,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,10240,32,0.0028808888875775864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,10240,32,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,16384,0.07141155666775174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,65536,0.2660328812069363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,16384,0.05335200164053175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,65536,0.3536355495452881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,16384,0.08784977595011394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,16384,0.06742311186260648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,12288,0.0395688878165351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,65536,0.19723378287421334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,12288,0.06718933582305908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,65536,0.23269245359632704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,12288,0.04608088731765747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,10240,0.05642844571007622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,10240,0.03369599911901686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,8192,0.045781334241231285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,8192,0.027969777584075928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,10240,0.04137688875198364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,8192,0.032815112007988825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,7168,0.02511022157139248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,7168,0.040531555811564125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,7168,0.030626667870415583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,6144,0.0358168880144755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,6144,0.02256533337963952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,5120,0.03005511230892605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,6144,0.02626577847533756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,5120,0.01964000033007728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,5120,0.023951111568344965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,4096,0.024615110622511968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,4096,0.01664977769056956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,3584,0.022031999296612207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,4096,0.019892444213231403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,3584,0.015574221809705099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,3072,0.01942488882276747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,3584,0.017986666825082567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,3072,0.013963555296262106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,3072,0.016390222642156813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,2560,0.01679111189312405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,2560,0.014803555276658801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,2048,0.013859555953078799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,2048,0.011204444699817233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,2048,0.01290844463639789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,1536,0.009886222581068674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,1536,0.011218667030334473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,1024,0.008088889221350351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,1536,0.011404444773991903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,1024,0.008397333323955536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,1024,0.010073777702119615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,768,0.007085333267847697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,768,0.007074666519959767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,768,0.008254222571849823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,512,0.00518844442235099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,512,0.005731555736727185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,512,0.006773333168692059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,256,0.0033644443998734155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,256,0.004693333473470476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,256,0.00619555554456181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,128,0.0042568888101312844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,128,0.003050666716363695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,64,0.0027884443600972495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,128,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,8192,32,0.0028284444577164124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,64,0.0058408888677755995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,8192,32,0.005843555761708154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,65536,0.30468445354037815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,12288,2560,0.01577422188387977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,16384,0.04680266645219591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,65536,0.17739555570814344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,16384,0.08072266976038615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,12288,0.03727822171317206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,16384,0.05316355493333605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,12288,0.06288355588912964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,12288,0.038578665918774076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,10240,0.031874666611353554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,65536,0.20424444145626494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,10240,0.05304533243179321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,8192,0.04349955585267809
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,10240,0.033137778441111244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,8192,0.029305779271655615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,8192,0.027519999278916255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,7168,0.039107554488711886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,6144,0.034753776258892484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,7168,0.024915556112925213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,6144,0.020968000094095867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,5120,0.029639111624823675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,5120,0.01831377711560991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,6144,0.022413333257039387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,5120,0.019865777757432725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,4096,0.022295999858114455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,4096,0.015564444992277356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,3584,0.019914666811625164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,4096,0.01644000079896715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,3584,0.014634667171372307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,3584,0.015336889359686108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,3072,0.013129777378506131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,3072,0.017831110292010836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,3072,0.014044443766276041
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,2560,0.015364444918102689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,2560,0.012063110868136087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,2560,0.012902222573757172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,2048,0.01274577776590983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,2048,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,2048,0.011487111449241638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,1536,0.010460444622569613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,1536,0.009383111364311641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,1536,0.01014755583471722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,1024,0.007856888903511895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,1024,0.007692444655630324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,1024,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,768,0.005744888964626525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,768,0.006752000086837345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,768,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,512,0.005392888767851724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,512,0.006160000132189856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,512,0.0052328887912962176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,256,0.003293333368168937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,256,0.005664000080691443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,256,0.00462400002612008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,128,0.003000000077817175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,128,0.005406222409672207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,128,0.0041164445380369825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,64,0.0027679999669392905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,7168,32,0.002765333279967308
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,64,0.005380444642570283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,7168,32,0.00535377777285046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,65536,0.26489422056410045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,16384,0.04741244514783224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,65536,0.14822222126854792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,65536,0.1619066662258572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,16384,0.07281066974004109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,12288,0.054795556598239474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,12288,0.032551109790802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,16384,0.04390933447413974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,12288,0.03496977686882019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,10240,0.04604355494181315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,10240,0.02805244591501024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,8192,0.03773244553142124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,8192,0.023224888576401606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,10240,0.030206223328908283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,7168,0.033398220936457314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,8192,0.023856000767813787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,7168,0.021180444293551978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,6144,0.027094221777386133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,6144,0.029254221253924902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,7168,0.02241777711444431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,6144,0.019280888968043856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,6144,0.020671111014154222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,5120,0.02458133300145467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,5120,0.017097777790493436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,3072,0.016328000360065036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,5120,0.01827022267712487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,4096,0.020640888147883944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,4096,0.014536889062987434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,3584,0.018557333283954196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,4096,0.01552977826860216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,3584,0.014092443717850579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,10240,2560,0.014711999230914645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,3072,0.016360888878504436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,3584,0.014386667145623101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,3072,0.012448888685968188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,3072,0.013282666603724161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,2560,0.014151111245155334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,2560,0.01205599970287747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,2048,0.01203555530971951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,2048,0.009866666462686326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,2048,0.01091733326514562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,1536,0.009709333380063375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,1536,0.009676444033781687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,1536,0.009335111412737105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,1024,0.006866666591829724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,1024,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,1024,0.007749333149856991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,768,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,768,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,768,0.006432000133726332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,512,0.005220444252093633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,512,0.004278222305907143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,512,0.006048888795905643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,256,0.0032551110618644287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,256,0.004813333352406819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,256,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,128,0.002983111060327954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,128,0.004314666820896996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,128,0.005321777943107817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,64,0.002756444530354606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,6144,32,0.002740444408522712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,64,0.005342222336265776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,6144,32,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,16384,0.03913244605064392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,65536,0.22403022978040907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,16384,0.05737511316935221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,12288,0.030237333642111883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,16384,0.04200177722507053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,65536,0.14077511098649767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,12288,0.04417333337995741
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,65536,0.12662933932410345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,12288,0.029137776957617864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,10240,0.03736977775891622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,10240,0.025493333737055462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,8192,0.030392888519499037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,8192,0.022245332598686218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,8192,0.021340444684028625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,7168,0.027119111683633592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,7168,0.020463110672103036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,6144,0.02405866649415758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,7168,0.019465777609083388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,6144,0.01808444493346744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,6144,0.01777066621515486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,5120,0.020416888925764296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,5120,0.015782222151756287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,4096,0.016964443855815463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,5120,0.015983111328548856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,4096,0.013580444786283704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,3584,0.015273777974976433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,4096,0.013583111266295115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,3584,0.01315199997689989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,3584,0.012658666405412885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,3072,0.013588444226317935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,3072,0.011468444433477191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,2560,0.011785777906576792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,3072,0.011800888511869641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,2560,0.011143999795118967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,2560,0.010977778169843884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,2048,0.010106667048401302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,2048,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,2048,0.010116444693671333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,1536,0.008829333715968663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,1536,0.008451555338170793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,1536,0.009131555755933126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,1024,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,1024,0.005821333163314396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,1024,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,768,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,768,0.005646222167544895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,768,0.006227555374304454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,512,0.005042666776312722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,512,0.00370399985048506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,512,0.005863111052248213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,256,0.003101333354910215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,256,0.005496888938877318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,256,0.004578666554556953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,128,0.004146666576464971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,128,0.0029120000286234748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,128,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,64,0.0027217778066794076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,5120,32,0.0027475555737813315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,32,0.005221333354711533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,5120,64,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,65536,0.18411289321051705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,65536,0.10876888699001735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,16384,0.04820000131924947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,16384,0.031173331869973078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,65536,0.11718932787577312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,12288,0.026857778429985046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,8192,2560,0.012770666844315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,12288,0.03826133410135905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,10240,0.03235822253757053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,10240,0.02350666622320811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,12288,0.02527911133236355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,10240,0.022050667140218947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,8192,0.02683111031850179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,8192,0.020406223005718656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,8192,0.01833866702185737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,7168,0.01869244376818339
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,7168,0.024301333559883967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,6144,0.021669333179791767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,7168,0.017064889272054035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,6144,0.0163102216190762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,5120,0.01907111042075687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,6144,0.01575199928548601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,5120,0.014892444014549255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,4096,0.015970667203267414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,5120,0.014202665951516895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,4096,0.012861332959598966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,3584,0.012841777669058906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,4096,0.012242666549152799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,3584,0.013047999805874295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,3584,0.011578666667143503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,3072,0.011500444677140979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,3072,0.011201777391963534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,3072,0.01072355525361167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,2560,0.010178666975763109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,2560,0.010110222631030613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,2560,0.010854221880435944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,2048,0.008758222063382467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,2048,0.008652444514963362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,2048,0.009338666995366415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,1536,0.0074684446056683855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,1536,0.008088000118732452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,1536,0.007351111206743453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,1024,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,1024,0.005877333382765452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,1024,0.006121777825885349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,768,0.004687110996908612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,768,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,768,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,512,0.0034497777620951333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,512,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,512,0.005112888912359874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,256,0.003076444483465619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,256,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,256,0.004553777890072929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,128,0.0041964443193541635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,128,0.002855111120475663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,128,0.005118222286303838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,64,0.0026444445053736367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,64,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,4096,32,0.0026435556097163092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,4096,32,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,65536,0.1554515494240655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,16384,0.03214755654335022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,65536,0.10796266794204712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,65536,0.09545422262615627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,16384,0.04324355390336779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,12288,0.033732444047927856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,12288,0.024932444095611572
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,16384,0.02846755584081014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,10240,0.029289778735902574
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,12288,0.023390221926901076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,7168,7168,0.02380977736579047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,10240,0.020599111914634705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,10240,0.022330666581789654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,7168,0.022219555245505437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,8192,0.01717422240310245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,8192,0.01868888901339637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,7168,0.01679022279050615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,8192,0.024326221810446844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,7168,0.016050666570663452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,6144,0.01958755486541324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,6144,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,5120,0.013880000346236758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,5120,0.017252445220947266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,5120,0.013517333401574029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,4096,0.012835555606418185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,4096,0.011984888878133563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,4096,0.011595555477672152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,3584,0.012040889097584618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,3584,0.011730666789743634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,3584,0.011056000159846412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,3072,0.010499555203649731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,3072,0.010433778166770935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,3072,0.01036977767944336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,2560,0.009365333451165093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,2560,0.010348444183667501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,2560,0.00998399986161126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,2048,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,2048,0.008181333541870117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,2048,0.008799110849698385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,1536,0.0070293330483966405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,1536,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,1536,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,1024,0.005741333381997214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,1024,0.005452444569932089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,1024,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,768,0.0037466668420367767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,768,0.00544177782204416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,768,0.005859555469618903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,512,0.0034248888906505373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,512,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,512,0.004920000003443824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,256,0.003050666716363695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,256,0.005386666705211003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,128,0.004106666478845808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,128,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,128,0.005160888863934411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,64,0.0026266665922270883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,64,0.005136888888147143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3584,32,0.0026400000270869997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3584,32,0.005135111096832487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,65536,0.13809866375393337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,65536,0.09538844558927749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,16384,0.030509332815806072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,16384,0.0385537776682112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,65536,0.09180000093248154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,16384,0.026400888959566753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,12288,0.023903111616770428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,12288,0.03035199973318312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,12288,0.02163644466135237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,10240,0.02633511192268795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,10240,0.021301334102948506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,10240,0.019075555933846366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,8192,0.017789334058761597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,8192,0.0219217770629459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,8192,0.01611911091539595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,7168,0.01628266606065962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,7168,0.019871999820073444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,7168,0.015080889066060385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,6144,0.01763199932045407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,6144,0.015007111761305066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,6144,0.013983999689420065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,5120,0.013568888935777875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,5120,0.015512888630231222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,5120,0.012897777888509961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,4096,0.011343110766675739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,4096,0.01401422255569034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,4096,0.011180444724029966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,3584,0.012929778132173749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,3584,0.011337777806652917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,3584,0.010571555958853828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,3072,0.009822222093741098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,3072,0.011702222128709158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,3072,0.010009778042634329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,2560,0.00982844415638182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,2560,0.008752889103359645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,2560,0.009376000199053023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,2048,0.007490666376219855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,2048,0.007619555625650618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,2048,0.007899555895063613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,1536,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,1536,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,1536,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,1024,0.004883555488453971
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,1024,0.005447111195988125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,1024,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,768,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,768,0.0037333332002162933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,768,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,512,0.004791999856630961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,512,0.0034302222645945023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,512,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,6144,2560,0.011944000091817645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,256,0.004412444515360726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,256,0.003020444470975134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,256,0.005398222141795688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,128,0.0027626665929953256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3072,128,0.004064889003833135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,128,0.005220444252093633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,64,0.002612444468670421
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,64,0.005175999883148405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,3072,32,0.0026604444202449587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,3072,32,0.005139555368158552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,65536,0.11476266384124756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,16384,0.028456889920764502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,16384,0.032538665665520564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,65536,0.07759110795127021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,16384,0.024359999431504145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,65536,0.0783368878894382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,12288,0.02622933354642656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,12288,0.02000799940692054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,10240,0.02253155575858222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,10240,0.019727110862731934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,10240,0.017828444639841717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,8192,0.019039111004935372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,8192,0.015196444259749519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,7168,0.015136000182893542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,7168,0.01741066740618812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,7168,0.014262222581439547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,6144,0.013946667313575745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,6144,0.015631111131774057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,6144,0.013254221942689685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,5120,0.0125928885406918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,5120,0.01385155568520228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,5120,0.012192000117566852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,5120,10240,0.026368000441127356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,4096,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,4096,0.01218311074707243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,4096,0.010685332947307162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,3584,0.010888888604111142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,3584,0.011378666592968835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,3584,0.010208889014191097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,3072,0.01055022246307797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,3072,0.009294222626421187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,3072,0.009671111073758867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,2560,0.008055999875068665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,2560,0.008611555728647444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,2560,0.008353778057628209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,2048,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,2048,0.0070773329999711775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,2048,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,1536,0.0063991112013657885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,1536,0.00619555554456181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,1536,0.0064239998658498125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,1024,0.004117333226733738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,1024,0.0053582220441765254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,1024,0.00600888869828648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,768,0.0037048889531029593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,768,0.005792888916201062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,512,0.0047351109484831495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,512,0.003391999958289994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,512,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,4096,16384,0.03327644533581204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,256,0.003089777711364958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,256,0.005378666851255629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,128,0.0027742222365405825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,128,0.005203555441564984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,64,0.0026382222357723448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,64,0.005164444446563721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2560,32,0.002640888922744327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2560,32,0.005187555733654234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,65536,0.08568533261617024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,65536,0.09334488709767659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,65536,0.07229599687788221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,16384,0.028406222661336262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,16384,0.026922666364245947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,16384,0.02274844381544325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,12288,0.023379556006855432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,12288,0.021557332740889654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,12288,0.01862044466866387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,6144,0.015253333581818474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,10240,0.02055022286044227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,10240,0.018948444061809115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,10240,0.016688888271649677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,8192,0.015841777125994366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,8192,0.014341332846217685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,7168,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,3584,256,0.004447111239035924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,7168,0.014635556273990206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,7168,0.0135057775510682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,6144,0.014310222533014087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,6144,0.013248888982666863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,6144,0.01257066677014033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,5120,0.013339555925793119
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,5120,0.012392000191741519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,5120,0.011607999602953592
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,4096,0.010843555960390302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,4096,0.011383111278216044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,4096,0.010112889111042023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,3584,0.011410666836632622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,3584,0.01016088906261656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,3584,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,3072,0.009538667069541084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,3072,0.008488888541857401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,2560,0.008949333594905006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,2560,0.00756444450881746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,2560,0.007394666473070781
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,2048,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,2048,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,2048,0.006726222319735422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,1536,0.007044444481531779
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,1536,0.005233777893914117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,1536,0.006215999937719769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,1024,0.005630222045713001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,1024,0.003926222109132343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,1024,0.005836444596449534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,768,0.005504000104135937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,768,0.003630222131808599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,768,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,512,0.004923555586073133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,512,0.0032746667663256326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,512,0.00544177782204416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,256,0.004442666553788715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,256,0.002934222213096089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,256,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,128,0.004150222159094281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,128,0.0027182222240500977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,128,0.00509688879052798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,64,0.002586666701568498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,12288,0.022616000639067754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,64,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,2048,32,0.0026293332791990707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,2048,32,0.005138666679461797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,8192,0.016858667135238647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,65536,0.07648711072074042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,65536,0.07016888591978285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,16384,0.025501334004931982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,16384,0.021884444687101576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,16384,0.021554667088720534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,65536,0.08361422353320652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,12288,0.020580444071027968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,12288,0.017885333961910672
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,12288,0.02164622147878011
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,10240,0.01576266686121623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,10240,0.01856355534659492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,8192,0.013554666605260638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,10240,0.019535111056433786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,8192,0.015538665983412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,8192,0.013743110828929476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,7168,0.012601777911186218
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,7168,0.01661688917213016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,7168,0.013010666602187686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,6144,0.012342222034931183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,6144,0.013187555803192986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,6144,0.012146666646003723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,5120,0.010528888967302112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,5120,0.011367110742463006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,4096,0.009422221945391761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,4096,0.009464888936943477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,4096,0.009276444713274637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,3584,0.009375111096435124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,3584,0.009043555292818282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,3584,0.008835555778609382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,3072,0.008552889029184977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,768,0.0052435555391841466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,3072,0.007867555651399825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,3072,0.007870222131411234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,2560,0.007062222394678328
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,2560,0.007126222054163615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,2048,0.006140444427728653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,2048,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,2048,0.006672888994216919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,1536,0.0046533333758513136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,1536,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,1536,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,1024,0.0053671110007498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,1024,0.003954666770166821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,1024,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,768,0.005140444470776452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,768,0.0036106668412685394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,768,0.005719999886221356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,512,0.0032666667054096856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,512,0.004726222405831019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,512,0.005480888817045424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,256,0.002966222249799304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,256,0.004353777815898259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,256,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,128,0.004058666527271271
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,128,0.002728888971938027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,128,0.005147555636035072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,64,0.0025804444319672054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,256,0.00433688900536961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1536,32,0.0025840000145965153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,64,0.00517866677708096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1536,32,0.005158222383923001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,65536,0.04876533481809828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,65536,0.06621155473921034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,16384,0.020670221911536325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,65536,0.0651822222603692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,16384,0.016393777396943834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,16384,0.021195555726687115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,12288,0.016972444123691983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,12288,0.013600000076823764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,12288,0.017237333787812125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,10240,0.012296888563368054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,10240,0.01575199928548601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,10240,0.01568888955646091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,8192,0.010802666346232096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,8192,0.013432888521088494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,7168,0.01201600001917945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,7168,0.010271111296282874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,8192,0.013461333182122974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,7168,0.012572444147533841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,6144,0.009279111193286048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,6144,0.010804444551467896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,6144,0.011476444701353708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,5120,0.00887022250228458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,5120,0.00923555592695872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,5120,0.00980977796845966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,4096,0.007836444510353936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,4096,0.008369777765538957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,4096,0.008199111455016667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,3584,0.008072000410821702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,3584,0.007816889219813878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,3584,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,3072,0.007224000162548489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,3072,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,3072,0.007481777833567725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,2560,0.0063982220987478895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,2560,0.00702044450574451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,2560,0.007052444749408298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,2048,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,2048,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,2048,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,1536,0.0058044443527857465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,1536,0.004603555632962121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,1536,0.006200888918505774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,1024,0.005173333403136996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,1024,0.003943110919660992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2560,128,0.004091555459631814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,1024,0.005848888721730974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,768,0.00360977773865064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,768,0.004800888813204235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,768,0.00565244464410676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,512,0.004605333424276776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,512,0.0032666667054096856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,512,0.005398222141795688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,256,0.0029688889367712867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,256,0.005231110999981563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,128,0.004019555532270008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,128,0.0027155555370781156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,64,0.0029111111329661477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,128,0.00509066672788726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,64,0.005102222164471944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,1024,32,0.002943111169669363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,1024,32,0.00508088908261723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,65536,0.038183109627829656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,65536,0.06538044744067721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,65536,0.06507733133104113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,16384,0.013966222604115805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,16384,0.020796444680955675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,12288,0.016190222567982145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,12288,0.011771555576059552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,12288,0.017109332813156974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,10240,0.014456889695591398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,10240,0.010822222464614444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,10240,0.015431111057599386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,8192,0.01219911128282547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,8192,0.00977244476477305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,8192,0.013075555364290873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,7168,0.010798222488827176
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,7168,0.009100444614887238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,7168,0.011238222320874533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,6144,0.00960355583164427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,6144,0.008434666527642144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,6144,0.01019200020366245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,5120,0.00868888861603207
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,5120,0.008178667061858708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,5120,0.009499555660618676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,4096,0.007788444558779399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,4096,0.007104000283612146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,4096,0.008244444098737504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,3584,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,3584,0.007060444189442529
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,3584,0.007877333296669854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,3072,0.00685777763525645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,3072,0.006534222099516127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,3072,0.007495111061467065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,2560,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,2560,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,2560,0.007079111205206976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,2048,0.005984888722499211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,2048,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,2048,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,1536,0.005479111025730769
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,1536,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,1536,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,1024,0.005096000101831225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,1024,0.003915555361244413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,1024,0.0058577776783042485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,8192,0.0177511109246148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,768,0.003552888830502828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,768,0.004729777574539185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,768,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,512,0.004524444540341695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,512,0.003257777748836411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,512,0.005445333404673471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,256,0.003169777699642711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,256,0.004122666600677702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,256,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,128,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,128,0.003944000022278892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,128,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,64,0.0027155555370781156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,64,0.005103999955786599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,768,32,0.0027359999302360746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,768,32,0.005142222262091107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,65536,0.027523555689387854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,65536,0.06487733125686646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,65536,0.06509244441986084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,16384,0.020042666130595736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,16384,0.010964444114102257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,2048,3072,0.009941333697901832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,16384,0.020821332931518555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,12288,0.01565511193540361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,12288,0.009760888914267222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,12288,0.016959110895792644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,10240,0.009162666896979014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,10240,0.013882666826248169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,10240,0.013556444810496436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,8192,0.00851733320289188
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,8192,0.011434666812419891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,8192,0.011696000066068439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,7168,0.010417777631017896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,7168,0.007789333661397298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,7168,0.01090755561987559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,6144,0.0070364442136552595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,6144,0.009531555904282464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,6144,0.010110222631030613
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,5120,0.006904888898134232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,5120,0.00868266655339135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,5120,0.009325332939624786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,4096,0.007724444071451823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,4096,0.006154666758245892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,4096,0.008277333445019191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,3584,0.006697777658700943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,3584,0.007266666326257918
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,3584,0.007898666792445714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,3072,0.0063182223174307085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,3072,0.0068400001360310455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,3072,0.0075039996041191956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,2560,0.006035555568006303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,2560,0.007088888850477006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,2560,0.0063368889192740125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,2048,0.0053484443989064955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,2048,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,2048,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,1536,0.0046844445168972015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,1536,0.005480000128348668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,1536,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,1024,0.004060444318585926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,1024,0.005024888863166173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,1024,0.005848888721730974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,768,0.0036844445599450003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,768,0.004919111314747068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,768,0.005660444498062134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,512,0.004405333350102107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,512,0.003346666693687439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,512,0.005424889011515512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,256,0.003010666618744532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,256,0.004067555483844546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,256,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,128,0.002761777697337998
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,512,128,0.003964444415436851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,128,0.005141333159473208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,64,0.0026293332791990707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,64,0.00510044437315729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,512,32,0.0026373333401150177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,512,32,0.005095110999213325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,65536,0.0169831116994222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,65536,0.06496622165044148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,16384,0.018985778093338013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,65536,0.06496977806091309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,16384,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,16384,0.018947554959191214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,12288,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,12288,0.015004444453451367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,12288,0.01497511068979899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,10240,0.013314666847387949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,10240,0.006739555547634761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,10240,0.013423999978436364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,8192,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,8192,0.011433777709801992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,8192,0.011583999627166323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,7168,0.006351999938488007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,7168,0.010481778118345471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,7168,0.010825778047243753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,6144,0.006145777801672618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,6144,0.00960888879166709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,6144,0.010002666877375709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,5120,0.006486222147941589
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,5120,0.009185777770148383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,5120,0.008598222500748104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,4096,0.00775111135509279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,4096,0.006063999815119638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,4096,0.00811288919713762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,3584,0.00654666663871871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,3584,0.007298666569921706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,3584,0.007744000189834171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,3072,0.006207110981146495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,3072,0.006804444309737947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,3072,0.007391999993059371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,2560,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,2560,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,2560,0.006990222467316522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,2048,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,2048,0.005223999834722943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,2048,0.006554666492674086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,1536,0.005535111245181825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,1536,0.004570666700601578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,1536,0.006135111053784688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,1024,0.005045333256324132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,1024,0.005752000129885144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,768,0.003581333491537306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,768,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,768,0.004625777817434735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,512,0.003256888853179084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,512,0.00443466669983334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,512,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,256,0.0029128889242808023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,256,0.004064889003833135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,256,0.005139555368158552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,128,0.002687999978661537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,256,128,0.003929777691761653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,128,0.005015999906592899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,64,0.0025555555605226094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,64,0.005048000150256687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,256,32,0.0025902222841978073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,256,32,0.005036444299750858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,65536,0.012061333490742577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,65536,0.0645937787161933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,16384,0.018703111343913607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,65536,0.0637635588645935
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,16384,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,16384,0.018223111828168232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,12288,0.01496266656451755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,12288,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,12288,0.014859555496109856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,10240,0.006383111079533895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,10240,0.013316444224781461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,8192,0.01129955550034841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,8192,0.006184000107977126
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,8192,0.011477332976129321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,7168,0.010395555860466426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,7168,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,7168,0.010679999987284342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,6144,0.009468444519572789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,6144,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,6144,0.009930666950013902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,5120,0.008591111335489485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,5120,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,5120,0.009154666629102495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,4096,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,4096,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,4096,0.008124444219801161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,3584,0.007182222273614671
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,3584,0.0063475556671619415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,3584,0.00776800016562144
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,3072,0.006057777752478917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,3072,0.007355555891990662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,2560,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,2560,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,2560,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,2048,0.005839111076460944
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,2048,0.005028444445795483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,2048,0.006531555619504716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,1536,0.005385777602593104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,1536,0.004413333204057482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,1536,0.006160000132189856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,1024,0.004950222041871813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,1024,0.003721777763631609
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,1024,0.005762666463851929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,768,0.004557333472702238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,768,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,768,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,512,0.004297777596447202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,512,0.003248888792263137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,512,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,256,0.003979555434650845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,256,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,256,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,128,0.0038240001433425476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,128,0.0027102221631341507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,128,0.005051555732885997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,64,0.002571555475393931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,5120,0.011876444849703046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,64,0.005044444567627377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,128,32,0.005027555757098728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,128,32,0.002604444407754474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,65536,0.00962933318482505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,16384,0.005988444305128521
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,16384,0.018325333793958027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,65536,0.06122844749026828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,12288,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,10240,0.006179555422729916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,12288,0.014946666028764514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,10240,0.01332355539004008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,8192,0.006089777582221561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,8192,0.011517333487669626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,7168,0.00601333338353369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,6144,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,7168,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,6144,0.009974222216341231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,5120,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,5120,0.009184888667530483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,4096,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,4096,0.008113777471913232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1536,2560,0.008095111283991072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,3584,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,3584,0.007717333734035492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,3072,0.006000888844331105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,2560,0.005519111123349931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,3072,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,2560,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,2048,0.00647555540005366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,2048,0.004929777648713854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,1536,0.006132444573773279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,1024,0.003714666598372989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,1536,0.004354666504595014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,1024,0.005753777921199799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,768,0.0034355556385384668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,768,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,512,0.003145777723855443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,512,0.005342222336265776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,256,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,256,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,128,0.00272444449365139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,128,0.005011555635266834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,64,0.002581333327624533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,64,0.005030222237110138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,64,32,0.002562666725781229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,64,32,0.005054222212897407
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,65536,0.008228444390826756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,16384,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,16384,0.01834133267402649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,12288,0.006116444451941385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,65536,0.05903466542561849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,12288,0.014953777194023132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,10240,0.006190222170617845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,10240,0.013376889129479727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,8192,0.006106666806671355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,8192,0.011495999991893768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,7168,0.010753778119881948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,7168,0.00600622221827507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,6144,0.005868444426192178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,6144,0.009972444838947719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,5120,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,4096,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,5120,0.009210666848553551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,4096,0.008112000094519721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,3584,0.006339555399285422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,3584,0.007752888732486301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,3072,0.00601244428091579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,3072,0.007330666813585494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,2560,0.005519111123349931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,2560,0.00700800038046307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,2048,0.004925333377387789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,1536,0.004315555509593752
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,2048,0.006545777950021956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,1536,0.006086222413513396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,1024,0.003689777933888965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,1024,0.005753777921199799
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,768,0.0034248888906505373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,768,0.005541333307822545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,512,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,512,0.005326222214433882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,256,0.0027928888383838865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,128,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,256,0.005166222237878376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,128,0.00500177757607566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,64,0.002528888897763358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,64,0.005034666508436203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,4,32,32,0.0025457777082920074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,4,32,32,0.005007110950019625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,16384,0.7126755714416504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,1024,256,0.004204444587230682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,16384,0.3768773343827989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,16384,0.4001928965250651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,12288,0.5417466693454319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,12288,0.31159289677937824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,768,16384,0.020038222273190815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,10240,0.4857653511895074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,8192,0.3608728779686822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,10240,0.25246577792697483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,10240,0.2738408777448866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,12288,0.2970924377441406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,8192,0.18840711646609834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,7168,0.31498667928907603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,8192,0.2200817796919081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,6144,0.27042312092251247
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,7168,0.18441332711113823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,6144,0.15953866640726724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,7168,0.18965066803826225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,5120,0.22455199559529623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,5120,0.11928355693817139
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,4096,0.18217778205871582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,6144,0.1697075499428643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,4096,0.09443022145165338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,5120,0.13188178009457058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,3584,0.1586880021625095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,4096,0.11125422848595513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,3584,0.08067555560006036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,3072,0.13639644781748453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,3584,0.09403733412424724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,2560,0.11398488945431179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,3072,0.07125066386328803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,2560,0.060425778230031334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,3072,0.07551378011703491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,2048,0.0908728904194302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,2560,0.06379377841949463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,1536,0.07012089093526204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,2048,0.052163554562462695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,2048,0.0522400008307563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,1024,0.04733688963784111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,1024,0.028011557128694322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,768,0.03704266746838888
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,1536,0.041005333264668785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,1024,0.03193333413865831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,768,0.022016889519161646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,512,0.0254737784465154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,768,0.025073778298166063
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,256,0.014300444059901767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,256,0.012144000165992312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,512,0.019734222027990553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,256,0.01661422186427646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,128,0.008793777889675563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,128,0.008908444808589088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,128,0.013781333135233985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,64,0.010679999987284342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,64,0.0081395556529363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,65536,32,0.010586666564146677
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,65536,32,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,65536,0.6998800171746148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,16384,0.10688800281948513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,65536,0.3746346632639567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,16384,0.17573955323961046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,65536,0.4334142208099365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,12288,0.14289155271318224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,12288,0.07122577561272515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,16384,0.11363377836015488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,10240,0.060621334446801074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,12288,0.0854391124513414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,10240,0.11845155556996663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,8192,0.10511200295554267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,8192,0.056368887424468994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,10240,0.06822222471237183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,7168,0.07827555470996432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,8192,0.05490933193100823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,7168,0.043655110730065234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,7168,0.049267556932237416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,6144,0.06794133451249865
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,6144,0.04319466816054451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,5120,0.05657688776652018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,5120,0.03312177790535821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,10240,0.013200889031092325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,4096,0.02755022214518653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,5120,0.03717422154214647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,4096,0.04594666759173075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,4096,0.030718220604790583
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,3584,0.040849778387281634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,3584,0.024842666255103216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,3072,0.03533422284656101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,3584,0.027975112199783325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,3072,0.022216889593336318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,2560,0.030142221185896132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,3072,0.024511110451486375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,2048,0.02460622290770213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,2048,0.016421332955360413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,2560,0.021720000439220007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,1536,0.019519999623298645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,2048,0.018760000665982563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,1536,0.01367377738157908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,1024,0.013861333330472311
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,1024,0.010936000280910067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,1536,0.015432000160217285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,1024,0.012712889247470431
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,768,0.011169777976142036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,768,0.00960533320903778
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,768,0.011225777367750803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,512,0.008360889222886827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,512,0.008238222036096785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,4,128,3072,0.006729777902364731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,512,0.009897777603732215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,256,0.005740444279379315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,256,0.004957333207130432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,256,0.007139555282062954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,128,0.003354666754603386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,128,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,64,0.003027555636233754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,128,0.004399999976158142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,64,0.006686222222116258
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,16384,32,0.003064888839920362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,16384,32,0.006544000158707301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,16384,0.07112889157401191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,16384,0.1363662216398451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,65536,0.5351164605882432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,16384,0.08357689115736221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,12288,0.06426578097873263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,12288,0.10176889101664226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,65536,0.3123679955800374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,12288,0.05955288807551066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,10240,0.04640977912478977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,10240,0.09441066450542873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,10240,0.050664888487921826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,8192,0.06823022498024835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,8192,0.038457777765062116
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,7168,0.06592177682452731
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,8192,0.04371111260520088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,7168,0.03422844409942627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,6144,0.05331110954284668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,6144,0.030247999562157526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,7168,0.03705599904060364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,5120,0.044439110491010875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,6144,0.032617777585983276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,5120,0.025969778498013813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,4096,0.03661511010593838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,5120,0.028407110108269587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,4096,0.021914665897687275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,4096,0.02344977855682373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,3584,0.019704888264338177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,1536,0.04085155659251743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,3584,0.031861331727769636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,3072,0.027785778045654297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,3584,0.021550221575631037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,3072,0.017818666166729398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,3072,0.019344889455371432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,2560,0.0236817780468199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,2560,0.015636444091796875
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,2048,0.019631110959582858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,2560,0.017133333616786532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,2048,0.015006222658687167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,1536,0.01147911118136512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,1536,0.015479111009173922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,65536,512,0.018696000178654987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,1536,0.012809777425395118
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,1024,0.009568889107969072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,1024,0.010954666468832227
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,1024,0.013017777767446307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,768,0.009552888572216034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,768,0.008416000339719985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,768,0.009902222288979424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,512,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,512,0.0074728892909155945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,512,0.008235555556085374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,256,0.004293333325121137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,256,0.0063973334100511335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,256,0.0047217777205838096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,128,0.003412444351447953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,128,0.005968889014588461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,128,0.004300444490379757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,64,0.0031271111220121384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,64,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,12288,32,0.006017777654859755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,12288,32,0.00309599998096625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,65536,0.43849600685967344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,16384,0.0711475544505649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,65536,0.24042932192484537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,65536,0.2664213445451525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,16384,0.112244447072347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,12288,0.08585155672497219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,12288,0.04791555470890469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,16384,0.06690933307011922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,12288,0.051444444391462535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,10240,0.07183644506666395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,10240,0.04103466537263658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,10240,0.04484000139766269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,8192,0.057667553424835205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,8192,0.033372445238961115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,8192,0.036914666493733726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,7168,0.051148444414138794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,7168,0.030233777231640283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,7168,0.03258400162061056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,6144,0.04467288984192742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,5120,0.037576887342664934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,6144,0.02696000039577484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,6144,0.029213332467608984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,5120,0.02360800074206458
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,4096,0.03050400151146783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,5120,0.025574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,4096,0.01962488889694214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,4096,0.021201777789327834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,3584,0.02718844513098399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,3584,0.01953866746690538
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,3072,0.02387199964788225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,3072,0.016174222032229107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,3072,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,2560,0.020319110817379422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,2560,0.014563555518786112
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,2048,0.016798221402698092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,2048,0.01254666679435306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,2560,0.015585778488053216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,1536,0.013274666335847644
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,2048,0.013933334085676404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,1536,0.010879110958841113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,1024,0.009827555881606208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,1536,0.012206222448084088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,1024,0.009170666337013245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,1024,0.010434666441546546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,768,0.008235555556085374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,768,0.008215111162927415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,768,0.00945688866906696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,512,0.006516444600290722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,512,0.005488889084921942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,512,0.007506666911972894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,256,0.003544888976547453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,256,0.004864888886610667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,256,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,128,0.003157333367400699
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,128,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,128,0.00443733317984475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,64,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,64,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,10240,32,0.0029120000286234748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,10240,32,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,6144,0.03851111067665948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,65536,0.35369332631429035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,16384,0.10470577743318345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,16384,2560,0.019374221563339233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,65536,0.19756089316474068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,65536,0.23137688636779785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,16384,0.059947556919521756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,12288,0.0675777792930603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,12288,0.03981688949796889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,12288,0.04627289043532478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,10240,0.05628266599443224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,10240,0.03384177883466085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,8192,0.04574755496448941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,10240,0.040644443697399564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,8192,0.02781244450145298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,7168,0.04056977894571092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,7168,0.025041777226659987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,8192,0.03396000133620368
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,6144,0.03568889035118951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,7168,0.03035822179582384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,6144,0.02240444388654497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,6144,0.02664533257484436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,5120,0.030045333835813735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,5120,0.023431999815834895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,4096,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,4096,0.024619556135601465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,4096,0.01956266661485036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,3584,0.015997333659066092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,3584,0.02193244463867611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,3584,0.018280888597170513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,3072,0.019351111518012155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,3072,0.014188443620999655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,3072,0.016456888781653512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,2560,0.012392889294359418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,2560,0.016717332932684157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,2048,0.010933332973056369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,2560,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,2048,0.013839999834696451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,2048,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,1536,0.009917333722114563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,1536,0.011250666446155973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,1536,0.011258666714032492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,1024,0.00830844458606508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,1024,0.008419555922349295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,1024,0.009718221922715506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,768,0.00665511108107037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,768,0.007152888509962294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,768,0.008224888808197444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,65536,0.26568354500664604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,512,0.005695999910434087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,512,0.005075555708673266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,256,0.0034506666577524612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,256,0.004615111069546806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,512,0.00683111117945777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,256,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,128,0.003040888864133093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,128,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,128,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,64,0.002805333377586471
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,64,0.005782222168313132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,8192,32,0.0028035555862718155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,8192,32,0.005864000154866113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,65536,0.3032746579911974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,16384,0.05384888913896349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,16384,0.08023111025492351
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,65536,0.17726755142211914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,65536,0.17830933464898002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,12288,0.062337776025136314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,16384,0.049231112003326416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,12288,0.03893688983387417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,12288,2048,0.013613333304723104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,10240,0.031187554200490315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,10240,0.05264177918434143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,10240,0.03284622232119242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,8192,0.04329599936803182
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,8192,0.027311111489931744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,7168,0.038562665383021034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,7168,0.023613333702087402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,6144,0.03428711162673102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,6144,0.021312889125612047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,7168,0.024800888366169397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,5120,0.029317332638634577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,6144,0.022092445029152766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,5120,0.018957333432303537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,5120,0.019634667370054457
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,4096,0.02239377796649933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,4096,0.01571644511487749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,4096,0.01631644368171692
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,3584,0.015086222026083203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,3584,0.01996888882584042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,3072,0.017705778280893963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,3584,0.015288000305493673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,3072,0.013373333546850415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,2560,0.015230221880806817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,3072,0.013990221752060784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,2560,0.011718221836619906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,2048,0.012771555946932899
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,2560,0.01272266689274046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,2048,0.010341333018408881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,2048,0.01145600030819575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,1536,0.01038933296998342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,1536,0.009291555318567488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,1024,0.007948444121413762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,1536,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,1024,0.008055999875068665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,1024,0.00851999968290329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,768,0.006765333314736684
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,768,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,512,0.005104889058404499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,512,0.005316444569163852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,512,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,256,0.004495999879307217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,256,0.003318222239613533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,256,0.005642666584915585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,128,0.004077333129114575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,128,0.002992000016901228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,128,0.00535377777285046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,64,0.0028008888992998335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,64,0.00535377777285046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,7168,32,0.002814222127199173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,7168,32,0.005285333428117964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,65536,0.15303911103142634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,65536,0.2651653289794922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,10240,3584,0.018064889642927382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,65536,0.1826204458872477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,16384,0.04084177811940511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,16384,0.07263288895289104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,16384,0.04124888777732849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,12288,0.05543377664354113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,12288,0.03226133187611898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,12288,0.03352888756328159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,16384,0.04956266615125868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,10240,0.027800000376171533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,10240,0.04559822214974297
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,8192,0.037653333610958524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,10240,0.030319998661677044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,8192,0.023177777727444965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,8192,0.02437155610985226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,7168,0.0332568883895874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,7168,0.021588444709777832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,6144,0.028941333293914795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,7168,0.022639999787012737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,6144,0.019468444916937087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,6144,0.02015466656949785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,5120,0.0247022228108512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,5120,0.017123555143674214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,5120,0.01833777791923947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,4096,0.020809777908855014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,4096,0.014319111903508505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,4096,0.01556711064444648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,3584,0.01866933372285631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,3584,0.01496266656451755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,3584,0.014188443620999655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,3072,0.01643022232585483
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,3072,0.012960000170601739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,3072,0.013225778109497495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,2560,0.014197332991494073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,2560,0.011433777709801992
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,2560,0.012200000385443369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,2048,0.012179555164443122
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,2048,0.010102222363154093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,2048,0.010832000109884473
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,1536,0.009848000274764167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,1536,0.009345778160625035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,1536,0.009709333380063375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,1024,0.007374222079912822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,1024,0.00776622196038564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,1024,0.0075288886825243635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,768,0.0063386667105886675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,768,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,768,0.0058462222417195635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,512,0.00509066672788726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,512,0.004509333521127701
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,512,0.006105777704053455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,256,0.0032640000184377036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,256,0.004593777573770947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,256,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,128,0.002974222310715251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,128,0.0053013331360287136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,64,0.005320888840489917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,6144,32,0.0028088889602157804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,6144,32,0.0052897776994440295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,65536,0.22396710183885363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,8192,5120,0.019719999697473314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,16384,0.045497778389188975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,65536,0.12812888622283936
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,65536,0.12860088878207737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,16384,0.057321780257754855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,16384,0.03789244426621331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,12288,0.044293334086736046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,12288,0.030579557021458943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,12288,0.02896089024013943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,10240,0.03713688916630215
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,10240,0.02555199960867564
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,10240,0.02896889050801595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,8192,0.021537777450349595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,8192,0.03035911255412632
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,7168,0.027049778236283198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,8192,0.021240888370407954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,7168,0.019351111518012155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,7168,0.02144711050722334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,6144,0.02382133404413859
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,6144,0.018067555295096505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,6144,0.017680888374646504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,5120,0.020474667350451153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,5120,0.015784000356992085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,4096,0.013414222333166333
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,4096,0.016925333274735343
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,4096,0.013543999857372709
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,3584,0.01389422184891171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,3584,0.015227556228637695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,3584,0.01274577776590983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,3072,0.012413333687517377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,3072,0.013600000076823764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,3072,0.01201333353916804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,2560,0.010647111468844943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,2560,0.011815999945004782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,2560,0.01108266661564509
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,2048,0.009476444787449306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,2048,0.010028444230556488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,2048,0.009941333697901832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,1536,0.008784000244405534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,1536,0.008405333591832055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,1536,0.009288888838556077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,1024,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,1024,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,1024,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,768,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,768,0.005763555566469829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,768,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,512,0.005032000028424793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,512,0.003657777690225177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,512,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,256,0.0031617778456873367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,256,0.004512000001139111
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,256,0.005493333356248007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,128,0.0029360000044107437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,128,0.004173333446184794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,64,0.0026951111439201566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,128,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,64,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,5120,32,0.0027315556589100095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,5120,32,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,65536,0.18367377916971842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,65536,0.11027377843856812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,16384,0.03314844436115689
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,16384,0.048208889034059316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,8192,0.02630399995379978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,12288,0.037045333120558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,65536,0.117031110657586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,12288,0.03794844282997979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,16384,0.03087644444571601
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,12288,0.025055110454559326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,10240,0.03235022226969401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,10240,0.02409066590997908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,10240,0.021853332718213398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,8192,0.02696177860101064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,8192,0.020229332976871066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,7168,0.024154666397306655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,8192,0.018360000517633226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,7168,0.019311111834314134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,6144,0.021687999367713928
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,7168,0.017049777838918898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,6144,0.01687733332316081
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,6144,0.015579554769727917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,5120,0.01905866629547543
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,5120,0.015290665957662793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,5120,0.01423111061255137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,4096,0.012846222354306115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,4096,0.015957333975368075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,4096,0.012136889000733694
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,3584,0.012825777961148156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,3584,0.01367111090156767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,3072,0.011524444652928246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,3584,0.011527111132939657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,3072,0.011948444777064852
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,3072,0.01072088877360026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,7168,768,0.0057760001056724125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,2560,0.010205333431561789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,2560,0.010048888623714447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,2560,0.010072000324726105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,2048,0.008844444321261512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,2048,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,1536,0.008214222060309516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,1536,0.007446222007274628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,1536,0.007918222082985772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,1024,0.005986666513813867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,1024,0.006443555570311016
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,1024,0.00618488879667388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,768,0.0046826667255825466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,768,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,768,0.005649777750174205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,512,0.003448888866437806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,512,0.004927111168702443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,512,0.005670222143332164
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,256,0.003064888839920362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,256,0.00535111129283905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,256,0.004488000025351842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,128,0.004156444635656145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,128,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,64,0.0026604444202449587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,128,0.005136888888147143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,64,0.005199999858935674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,4096,32,0.002640888922744327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,4096,32,0.005202666752868229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,65536,0.15474843978881836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,16384,0.03402488761478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,65536,0.10262755552927653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,65536,0.09461155864927505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,16384,0.04308088951640659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,12288,0.03362844387690226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,16384,0.028359111812379625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,12288,0.02478755513827006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,12288,0.023246222072177466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,10240,0.029213332467608984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,10240,0.022171555293930903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,8192,0.024150222539901733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,10240,0.02033066584004296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,8192,0.018584888842370775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,7168,0.02202133337656657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,8192,0.017070222232076857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,7168,0.01776355504989624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,7168,0.01605955594115787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,6144,0.015569777952300178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,6144,0.019502222537994385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,5120,0.017222222354676988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,6144,0.014771555860837301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,5120,0.014392889208263822
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,4096,0.011911999848153857
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,4096,0.012801777985360889
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,5120,0.013377777404255338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,3584,0.011725333001878528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,4096,0.01163733336660597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,3584,0.012832000023788877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,3072,0.010562666588359408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,3584,0.010999110837777456
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,3072,0.010870222416188983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,3072,0.010287111004193624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,2560,0.009359999663299983
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,2560,0.009735999835862054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,2560,0.009659555223253038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,2048,0.008215111162927415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,2048,0.008907555705971187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,2048,0.008036444584528605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,1536,0.007040888898902469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,1536,0.0074142225914531285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,1536,0.006702222343948152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,1024,0.006055999961164262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,1024,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,1024,0.00609422226746877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,768,0.004080888711743885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,768,0.005409777578380372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,768,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,512,0.0034355556385384668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,512,0.00490044429898262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,512,0.005655111124118169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,256,0.0030337776988744736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,256,0.004380444271696938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,256,0.005378666851255629
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,128,0.0028071111689011254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3584,128,0.004050666673315896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,128,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,64,0.0026231110095977783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,64,0.005192000004980299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3584,32,0.0026293332791990707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3584,32,0.005159111072619756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,65536,0.1379342211617364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,16384,0.03066133459409078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,16384,0.038134220573637224
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,65536,0.09746222363577949
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,65536,0.0859866672092014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,16384,0.02619733413060506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,12288,0.02381955583890279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,12288,0.029913776450686987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,12288,0.021547555923461914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,10240,0.02606311109330919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,10240,0.020564445190959506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,10240,0.018951111369662814
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,8192,0.02174577779240078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,8192,0.016192888220151264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,8192,0.017737777696715463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,7168,0.0196426659822464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,7168,0.016714667280515034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,7168,0.015075556106037565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,6144,0.01756088932355245
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,6144,0.014771555860837301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,6144,0.014012444350454541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,5120,0.015512888630231222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,5120,0.013298666311634911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,5120,0.01274311128589842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,4096,0.014128888646761576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,4096,0.011337777806652917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,4096,0.011222222612963783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,3584,0.012893333203262754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,3584,0.012055111428101858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,3584,0.010615111225181155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,3072,0.010363555616802638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,3072,0.011744000017642975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,3072,0.00996800015370051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,2560,0.00871022211180793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,2560,0.009191999832789103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,2560,0.009555555880069733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,2048,0.007701333198282454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,2048,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,2048,0.007983110845088959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,1536,0.00666133314371109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,1536,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,1536,0.006543111056089401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,1024,0.004776000148720211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,1024,0.006064888917737537
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,1024,0.00609422226746877
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,768,0.003960888832807541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,768,0.005372444374693765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,768,0.005864000154866113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,512,0.003415999934077263
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,512,0.004746666798988978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,512,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,256,0.003076444483465619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,256,0.0053386667536364655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,256,0.004335999902751711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,3072,128,0.004036444342798657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,128,0.002790222151411904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,64,0.0026515554636716843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,128,0.00518222235971027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,64,0.005140444470776452
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,3072,32,0.0026684444811609057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,3072,32,0.005162666655249066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,65536,0.11359555191463894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,16384,0.028010666370391846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,65536,0.0839831100569831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,65536,0.07799822092056274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,16384,0.03250666790538364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,16384,0.024136000209384497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,12288,0.026055110825432673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,12288,0.022054221895005967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,10240,0.02251022226280636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,12288,0.01996977792845832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,10240,0.018928888771269057
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,8192,0.01905777719285753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,10240,0.017910222212473553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,8192,0.016640888320075143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,6144,128,0.004276444514592488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,8192,0.015240889456537036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,7168,0.015432000160217285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,7168,0.017309332887331646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,7168,0.014184888866212634
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,6144,0.014093332820468478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,6144,0.015565334094895257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,6144,0.013229333692126803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,5120,0.013905778527259827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,5120,0.012122666670216454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,4096,0.012109333442317115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,4096,0.010823999842007955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,4096,0.010620444185203977
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,3584,0.011065777805116443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,3584,0.011354666617181567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,3584,0.010109333528412713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,3072,0.01054222219520145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,3072,0.009754666851626502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,3072,0.009639110830095079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,2560,0.008068444828192392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,2560,0.008476444416575963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,2560,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,2048,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,2048,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,2048,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,1536,0.006219555520349079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,1536,0.0061928890645504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,1536,0.0063768890168931745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,1024,0.0057964444988303715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,1024,0.0041519999504089355
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,1024,0.006034666465388403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,768,0.0037493333220481873
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,768,0.005818666683302985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,768,0.005197333378924264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,512,0.0033742222521040174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,512,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,512,0.004615111069546806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,256,0.0030782222747802734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,256,0.0053075556125905775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,256,0.004261333495378494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,128,0.003951111187537511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,128,0.002776888923512565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,64,0.002603555512097147
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,128,0.005148444324731827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,64,0.005144000053405762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2560,32,0.0026275554878844153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2560,32,0.005154666801293691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,65536,0.09294755591286553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,16384,0.02861422300338745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,65536,0.0865395532713996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,65536,0.07165955834918551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,16384,0.02682488825586107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,16384,0.02236888971593645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,12288,0.021528888079855178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,12288,0.023193778263198003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,10240,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,12288,0.018421333697107103
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,10240,0.02083555526203579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,5120,5120,0.016142222616407607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,10240,0.0165964447789722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,8192,0.0159244438012441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,8192,0.017616000440385606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,7168,0.01459377838505639
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,8192,0.01423377792040507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,7168,0.013364444176355997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,7168,0.017248888810475666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,6144,0.01330044451687071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,6144,0.01240888900227017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,5120,0.013800000150998434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,5120,0.012249777714411417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,5120,0.011550222006109027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,4096,0.010816888676749336
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,4096,0.011388444238238864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,4096,0.010086221827401055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,3584,0.010222222242090438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,3584,0.009661333428488838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,3072,0.009528888596428765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,3072,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,3072,0.008836444053384993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,2560,0.008237332933478886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,2560,0.00742222203148736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,2560,0.007118222614129384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,2048,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,2048,0.006884444504976273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,2048,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,1536,0.005211555709441503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,1536,0.006858666737874349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,1536,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,1024,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,1024,0.003950222084919612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,1024,0.005792888916201062
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,768,0.005488889084921942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,768,0.0036106668412685394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,768,0.005688888745175467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,512,0.003310222178697586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,512,0.00481599983241823
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,512,0.00544177782204416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,256,0.002970666728085942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,256,0.004309333446953031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,256,0.005220444252093633
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,128,0.004092444562249714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,128,0.0027280000762807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,128,0.005084444251325396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,64,0.0025964443468385274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,64,0.005137777576843898
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,2048,32,0.002576888849337896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,2048,32,0.005120000077618493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,65536,0.07021155622270372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,65536,0.0802124473783705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,16384,0.025434666209750708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,65536,0.06775289111667208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,16384,0.021746666895018682
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,16384,0.021382222572962444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,12288,0.021328000558747187
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,12288,0.01776266594727834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,12288,0.01755288905567593
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,10240,0.01772622267405192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,10240,0.015662221444977652
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,10240,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,8192,0.015332443846596612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,8192,0.013575110998418597
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,8192,0.013648000028398303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,7168,0.015359110302395292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,7168,0.012596444123321109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,12288,0.02678666677739885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,7168,0.013037333057986366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,6144,0.011635555161370171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,6144,0.01220444424284829
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,6144,0.012993777791659037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,5120,0.010655110908879174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,5120,0.012316444681750404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,5120,0.011177777416176267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,4096,0.009565333525339762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,4096,0.009726222190592024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,4096,0.009649777577983009
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,3584,0.00903555585278405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,3584,0.010149333212110732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,3584,0.008557333714432186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,3072,0.008587555752860175
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,3072,0.008592888712882996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,2560,0.007064888874689738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,3072,0.0074417781498697065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,2560,0.007178666690985362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,2048,0.005964444329341252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,2048,0.0064533332155810455
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,2048,0.006726222319735422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,1536,0.004611555486917496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,1536,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,1536,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,1024,0.003962666624122196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,1024,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,1024,0.0058257778485616045
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,768,0.003612444632583194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,768,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,512,0.003286222202910317
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,512,0.0046364445653226645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,512,0.005410666680998272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,256,0.004242666479614046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,256,0.002945777856641345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,256,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,128,0.0027137777457634607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,128,0.003986666599909465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,128,0.005138666679461797
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,64,0.002557333351837264
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,4096,2048,0.008993777963850234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,64,0.005129777722888523
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1536,32,0.005127999931573868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1536,32,0.002578666640652551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,65536,0.048419555028279625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,65536,0.06614844666586982
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,16384,0.0210479994614919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,65536,0.06566488742828369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,16384,0.01637066735161675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,16384,0.02085955606566535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,12288,0.016895999511082966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,12288,0.013627555635240344
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,12288,0.017160000072585214
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,10240,0.015545777148670621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,10240,0.012364444633324942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,10240,0.015592889653311836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,8192,0.013694222602579327
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,8192,0.010834666589895884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,8192,0.013395555317401886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,7168,0.010192888478438059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,7168,0.012605333493815528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,7168,0.01254666679435306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,6144,0.010544000400437249
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,6144,0.009305777649084726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,6144,0.011193777951929303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,5120,0.00888177752494812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,5120,0.008943999807039896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,5120,0.009619555539555019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,4096,0.007778666913509369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,4096,0.008290666672918532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,4096,0.008192000289758047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,3584,0.0075626663035816616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,3584,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,3072,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,3072,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,3072,0.007460444337791867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,2560,0.0063795554969045855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,2560,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,2560,0.007039111521508958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,2048,0.006020444548792309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,2048,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,2048,0.006633777585294511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,1536,0.005718222094906702
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,1536,0.004594666676388847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,1536,0.006194666855865055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,1024,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,1024,0.003921777837806278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,1024,0.0058240000572469496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,768,0.004820444517665439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,768,0.0036115555299652945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,768,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,512,0.004400889078776042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,512,0.003265777809752358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,512,0.005430222385459476
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,256,0.004109333372778363
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,256,0.0029271110478374693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,256,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,128,0.0039822223285834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,128,0.0026986667265494666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,128,0.005065777649482091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,64,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,64,0.005085333353943295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,1024,32,0.002850666642189026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,1024,32,0.005096000101831225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,65536,0.06553244590759277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,65536,0.03807199994723002
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,65536,0.06488444407780965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,16384,0.01371022231049008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,16384,0.02015555567211575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,16384,0.02075111038155026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,12288,0.016217778126398723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,12288,0.011759111450778114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,12288,0.017154667112562392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,10240,0.014273777604103088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,10240,0.010817777779367236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,10240,0.015556444724400839
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,8192,0.012250666817029318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,8192,0.009754666851626502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,8192,0.012994666894276937
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,7168,0.010811555716726514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,7168,0.009073778159088558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,7168,0.011176889141400656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,6144,0.009594666461149851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,6144,0.008384000096056197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,6144,0.010159999959998662
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,5120,0.008657777474986183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,5120,0.008221333225568136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,5120,0.009257777697510189
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,4096,0.007080888582600488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,4096,0.00777599960565567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,4096,0.008243555823961893
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,3584,0.006978666616810693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,3584,0.007711110843552484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,3584,0.007826666865083907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,3072,0.006882666713661618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,3072,0.0063573333124319715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,3072,0.007507555186748505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,2560,0.006228444476922353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,2560,0.005983110931184556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,2560,0.007132444116804335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,2048,0.006009777800904379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,2048,0.0053048887186580235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,2048,0.006667555620272954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,1536,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,1536,0.00462666650613149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,1536,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,1024,0.00510044437315729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,1024,0.003945777813593547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,1024,0.005858666780922148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,768,0.004791999856630961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,768,0.003605333467324575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,768,0.005672889037264719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,512,0.004461333155632019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,512,0.0032968889507982465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,512,0.005464000006516774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,256,0.004150222159094281
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,256,0.003173333282272021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,256,0.0052977779673205475
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,768,128,0.004002666721741359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,128,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,128,0.005163555757866966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,64,0.002712888850106133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,64,0.005158222383923001
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,768,32,0.002741333304180039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,768,32,0.005141333159473208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,65536,0.06564178069432576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,65536,0.027464888162083093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2560,5120,0.012461333639091916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,65536,0.06509955724080403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,16384,0.019833778341611225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,16384,0.010980444649855295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,16384,0.02067377832200792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,12288,0.015964445140626695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,12288,0.009645333720578088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,12288,0.016878222425778706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,10240,0.013332444760534497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,10240,0.009078222016493479
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,10240,0.013647110925780402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,8192,0.011360000405046674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,8192,0.0085235552655326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,8192,0.01163111130396525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,7168,0.010488889283604093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,7168,0.008042666647169325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,7168,0.010841777755154504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,6144,0.00702044450574451
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,6144,0.009551111194822524
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,6144,0.010072000324726105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,5120,0.008599110775523717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,5120,0.006828444285525217
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,5120,0.009280888570679558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,4096,0.007693332930405934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,4096,0.006168889088763132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,4096,0.008211555580298105
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,3584,0.007126222054163615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,3584,0.006678222368160884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,3584,0.007812444534566667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,3072,0.00676711110605134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,3072,0.0063368889192740125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,3072,0.0074622225430276655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,2560,0.006083555519580841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,2560,0.006173333360089197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,2560,0.007070221834712558
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,2048,0.005407111098368962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,2048,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,2048,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,1536,0.005346666607591841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,1536,0.004703999807437261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,1536,0.006214222146405115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,1024,0.004025777594910728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,1024,0.005818666683302985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,768,0.0037164443896876443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,768,0.004678222040335338
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,768,0.005650666852792104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,512,0.004294222427739037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,512,0.0033813334173626373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,512,0.005418666534953647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,256,0.0041253334946102565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,256,0.003006222140457895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,256,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,128,0.003922666526503033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,128,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,128,0.005121777868933148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,64,0.0026737778551048706
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,3584,0.012372444073359171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,64,0.005105777747101254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,512,32,0.0026462222966882917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,512,32,0.005110222018427319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,65536,0.016930666234758165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,65536,0.0648133357365926
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,65536,0.06452888912624784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,16384,0.018869333797030978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,2048,6144,0.014871110518773397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,16384,0.008382221890820397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,16384,0.01943466729587979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,12288,0.015004444453451367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,12288,0.007111111448870764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,12288,0.014850666125615438
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,10240,0.006782222125265334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,10240,0.013202666408485837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,10240,0.013374221821626028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,8192,0.011355555719799466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,8192,0.00646666685740153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,8192,0.011510222322411008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,7168,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,7168,0.010435555544164447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,7168,0.010743111371994019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,6144,0.006177777631415262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,6144,0.009958222508430481
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,5120,0.00654666663871871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,5120,0.008542222281297049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,5120,0.009157333109113906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,4096,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,4096,0.006107555495368109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,4096,0.008112000094519721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,3584,0.007076444725195567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,3584,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,3584,0.00777333312564426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,3072,0.00674222202764617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,3072,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,3072,0.007366221812036302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,2560,0.006147555592987273
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,2560,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,2560,0.0069982219073507525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,2048,0.0058782220714622075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,2048,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,2048,0.006523555351628198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,1536,0.0052915554907586836
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,1536,0.004598222259018156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,1536,0.006164444403515921
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,1024,0.004965333475006951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,1024,0.003923555629120933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,1024,0.0057608890864584185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,768,0.003584889074166616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,768,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,512,0.004355555607212914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,512,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,512,0.005351999981535806
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,256,0.004026666697528628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,256,0.002918222298224767
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,256,0.005147555636035072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,128,0.003891555385457145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,128,0.002704888789190186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,128,0.004998222407367495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,64,0.002578666640652551
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,64,0.005053333524200651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,256,32,0.002579555536309878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,256,32,0.005050666630268097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,65536,0.0645315580897861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,65536,0.011947555674446953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,65536,0.06306311157014635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,16384,0.01867466668287913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,16384,0.006125333408514659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,16384,0.018251554833518136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,12288,0.014959999256663851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,12288,0.006052444378534953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,12288,0.014870221416155497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,10240,0.013119999733236102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,10240,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,10240,0.01330133361948861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,2560,0.007373332977294922
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,8192,0.01126844435930252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,8192,0.006227555374304454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,8192,0.011518222590287527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,7168,0.006148444281684027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,7168,0.010454222559928894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,7168,0.010775999890433418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,6144,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1536,768,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,6144,0.009944000177913243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,5120,0.008585777547624376
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,5120,0.006330666856633292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,5120,0.009220444493823582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,4096,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,4096,0.007679111427730984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,4096,0.008117333054542542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,3584,0.006446222050322427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,3584,0.007745777567227681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,3072,0.006046222315894232
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,3072,0.007393777370452881
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,2560,0.006109333286682765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,2560,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,2560,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,2048,0.0058462222417195635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,2048,0.005008888741334279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,2048,0.006528888725572162
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,1536,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,1536,0.004400889078776042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,1536,0.006141333530346553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,1024,0.004881777697139316
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,1024,0.0037360000941488477
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,1024,0.005751111027267244
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,768,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,768,0.003572444534964032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,768,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,512,0.004230222354332606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,512,0.0032666667054096856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,512,0.005320888840489917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,256,0.0039813332259655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,1024,3584,0.00832711077398724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,256,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,256,0.005167110926575131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,128,0.0038355555799272326
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,128,0.0026977778308921387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,128,0.005038222091065513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,64,0.002546666603949335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,64,0.005019555489222209
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,128,32,0.00253866674999396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,128,32,0.005024000174469418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,65536,0.009465778039561378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,65536,0.06121066543791029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,16384,0.00611200018061532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,16384,0.01831377711560991
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,12288,0.006048888795905643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,12288,0.014956444501876831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,10240,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,10240,0.013363555901580386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,8192,0.006148444281684027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,8192,0.011541333463456897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,7168,0.006018666757477655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,7168,0.010763555765151978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,6144,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,6144,0.00997511131895913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,5120,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,5120,0.009175111022260454
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,4096,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,4096,0.00811822215716044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,3584,0.0063733334342638654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,3584,0.007720888488822513
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,3072,0.006034666465388403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,3072,0.007383999725182851
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,2560,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,2560,0.0069653333889113525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,2048,0.004917333523432414
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,2048,0.006498666687144174
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,1536,0.004344888859324985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,1536,0.006163555714819167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,1024,0.00370133337047365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,1024,0.0057546666098965546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,768,0.0034106667670938703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,768,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,512,0.0031022222505675424
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,512,0.0053226666318045715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,256,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,256,0.005130666825506422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,128,0.0026844443960322272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,128,0.004993777722120285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,64,0.002532444480392668
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,64,0.005053333524200651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,64,32,0.002540444541308615
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,64,32,0.005059555586841372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,65536,0.008151999778217739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,16384,0.00611200018061532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,65536,0.059578663773006864
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,16384,0.018351111147138808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,12288,0.006092444476154115
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,12288,0.01493866741657257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,10240,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,10240,0.013370667066839008
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,8192,0.006144000010357962
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,8192,0.011546666423479715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,7168,0.006031111296680238
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,6144,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,7168,0.010752889017264048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,6144,0.010016000105275048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,5120,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,5120,0.00925688859489229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,4096,0.005848000033034219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,4096,0.008182221816645728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,3584,0.0064604443808396654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,3584,0.008037333687146505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,3072,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,2560,0.005534222142563925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,3072,0.007418666448858049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,2560,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,2048,0.004965333475006951
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,2048,0.006494222415818109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,1536,0.004398222184843487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,1024,0.0037573331760035623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,1536,0.006131555471155379
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,768,0.0034088889757792154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,1024,0.0057751110030545135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,768,0.00552977787123786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,512,0.003136000078585413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,512,0.0053377776510185665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,256,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,256,0.005144888742102517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,128,0.0026400000270869997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,128,0.00500177757607566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,64,0.002570666579736604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,64,0.005034666508436203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,2,32,32,0.0025439999169773525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,2,32,32,0.005006222261322869
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,16384,0.7121857537163628
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,16384,0.37455201148986816
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,16384,0.3754524389902751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,12288,0.2891262107425266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,512,1024,0.005081777771313985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,12288,0.5842009120517307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,12288,0.29285422960917157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,6144,0.009541333549552495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,10240,0.4553430875142415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,256,768,0.004651555584536659
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,10240,0.2530684471130371
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,8192,0.36192088656955296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,8192,0.21024266878763834
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,7168,0.3134844568040636
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,7168,0.18378133243984648
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,8192,0.20507289303673637
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,7168,0.16148444016774496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,6144,0.2700631088680691
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,6144,0.15972710980309382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,5120,0.22428178787231445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,5120,0.11319288942548965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,6144,0.15969155894385442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,4096,0.18136000633239746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,5120,0.12681155734592015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,3584,0.007082666787836287
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,6144,0.009458666874302758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,4096,0.10756266779369777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,2,128,3072,0.00674044465025266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,4096,0.1203920046488444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,3584,0.1584764454099867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,3072,0.13627021842532688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,3584,0.08053600125842623
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,2560,0.11453955703311497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,3584,0.0898853341738383
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,3072,0.07119555605782403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,3072,0.07855821980370416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,2560,0.06277510854932997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,2048,0.09071111016803318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,2048,0.049237334065967135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,1536,0.07020533084869385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,2560,0.0699253347184923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,1536,0.03906044363975525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,1024,0.04710222283999125
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,2048,0.0604355533917745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,1024,0.027475555737813313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,768,0.03684711125161912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,1536,0.05713511175579495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,1024,0.03978044456905789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,768,0.021909332937664453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,768,0.03584266702334086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,512,0.025216889050271776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,512,0.017251556118329365
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,256,0.014339556296666464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,128,0.008728889127572378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,512,0.03142666816711426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,256,0.02684177789423201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,128,0.00887288898229599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,256,0.01407199932469262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,128,0.025088000628683303
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,64,0.019786667492654588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,64,0.00793511089351442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,65536,32,0.0064444442590077715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,65536,32,0.019367999500698514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,16384,0.17362222406599256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,16384,0.09117688735326131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,12288,0.07134311066733466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,12288,0.13235555754767522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,16384,0.10292977756924099
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,10240,0.10939644442664252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,10240,0.05964089102215237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,12288,0.07669599850972493
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,10240,0.06478399700588651
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,8192,0.04869066675504049
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,8192,0.08892622258928086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,65536,0.7028622097439237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,8192,0.053463111321131386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,7168,0.07786577939987183
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,7168,0.046104888121287026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,6144,0.06799733638763428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,7168,0.05229599939452278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,6144,0.04333688815434774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,5120,0.0565262238184611
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,6144,0.04557777775658501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,5120,0.0328746669822269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,4096,0.04600800077120463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,5120,0.03779733180999756
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,3584,0.0407315558857388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,3584,0.024869332710901897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,4096,0.038310223155551486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,3072,0.03560622202025519
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,3584,0.0295466681321462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,3072,0.030608889129426744
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,2560,0.030035555362701416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,65536,0.37438223097059464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,2560,0.019316444794336956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,2560,0.027927110592524212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,65536,0.39732000562879777
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,2048,0.016395555602179635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,2048,0.021143111917707656
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,2048,0.02624711063173082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,1536,0.020622221959961783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,1024,0.010797333386209277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,1024,0.014712888333532544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,1536,0.01832177738348643
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,768,0.009562667045328353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,768,0.014178666803571912
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,768,0.01185955521133211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,1024,0.015841777125994366
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,512,0.007997333175606197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,512,0.008943999807039896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,256,0.004922666483455234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,512,0.013080000049538083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,256,0.00906755526860555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,256,0.005983110931184556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,128,0.003362666608558761
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,128,0.004432888908518685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,64,0.003026666740576426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,128,0.008355555435021719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,16384,32,0.00304888892504904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,64,0.00851466672288047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,16384,32,0.008364444805516137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,65536,10240,0.21850044197506377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,16384,0.08394310871760051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,65536,0.5485359827677408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,16384,0.13559199704064265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,12288,0.05680711401833428
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,65536,0.290737787882487
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,16384,0.08620355526606242
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,12288,0.10402133067448933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,10240,0.08553955290052627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,65536,0.3009902106391059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,10240,0.046294222275416054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,12288,0.06221155325571696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,8192,0.07412978013356526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,8192,0.04458399944835239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,10240,0.05070399906900194
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,8192,0.04225333200560676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,7168,0.06082044707404243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,7168,0.033921778202056885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,7168,0.03760088814629449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,6144,0.05247822072770861
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,6144,0.030040889978408813
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,6144,0.03322044346067641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,5120,0.044433779186672635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,4096,0.03583822316593594
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,5120,0.029005332125557795
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,4096,0.021543999512990315
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,3584,0.03162844313515557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,3584,0.019741333193249173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,4096,0.024347555306222703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,3072,0.02762311034732395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,3584,0.02245866590076023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,3072,0.01772888832622104
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,2560,0.01548266741964552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,2560,0.02364088926050398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,3072,0.020418667131000094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,2048,0.019612444771660697
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,2560,0.01832711034350925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,2048,0.01367555558681488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,2048,0.01625777781009674
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,1536,0.01144266708029641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,1536,0.01756177842617035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,1536,0.01426755554146237
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,1024,0.011392888923486074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,1024,0.00941422250535753
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,1024,0.012116444607575735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,768,0.009848000274764167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,768,0.00831733312871721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,512,0.005774222314357758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,512,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,768,0.011188444164064197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,512,0.00849422232972251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,256,0.004181333300140169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,256,0.007191110816266801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,256,0.004671111288997862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,128,0.0042462220622433555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,128,0.0033271111961868075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,64,0.003097777772280905
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,128,0.006704888823959563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,12288,32,0.0030382221771611106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,64,0.006796444455782573
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,12288,32,0.006709333509206772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,16384,0.06177510817845663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,16384,0.11155733797285293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,65536,0.438460456000434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,12288,0.055050667789247304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,16384,0.06578755378723145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,12288,0.08559822373920017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,65536,0.24835289849175346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,65536,0.25860799683464897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,12288,0.05232266585032145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,10240,0.04067733221583896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,10240,0.07730844285753039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,8192,0.033167110549079046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,8192,0.05743200249142117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,10240,0.044547554519441396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,8192,0.03676266802681817
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,7168,0.050901333491007485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,7168,0.03005689051416185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,6144,0.04459733433193631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,7168,0.033575110965304904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,6144,0.026838221483760413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,5120,0.023281777898470562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,6144,0.030024889442655776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,5120,0.037414222955703735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,5120,0.026408889227443274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,4096,0.03046755658255683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,4096,0.019367999500698514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,4096,0.022408000297016565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,3584,0.027092445227834914
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,3584,0.017847999930381775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,3584,0.020626667473051283
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,3072,0.016214221715927124
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,3072,0.023766222927305434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,3072,0.018788443671332467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,2560,0.020297777321603563
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,2560,0.017151999804708693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,2048,0.012682666381200155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,2048,0.016733333468437195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,2048,0.015280889140235053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,1536,0.010778666370444827
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,1536,0.013363555901580386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,1536,0.01349688900841607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,1024,0.008960000342792934
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,1024,0.009889778163697984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,1024,0.011519999967681037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,768,0.008151111503442129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,768,0.01073155552148819
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,512,0.005412444472312927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,512,0.006487111250559489
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,512,0.007894222107198503
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,256,0.004782222211360931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,256,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,256,0.007069333559936947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,128,0.004367111043797599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,128,0.0031724443866146933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,128,0.006670222100284364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,64,0.0028737777223189673
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,64,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,10240,32,0.002855111120475663
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,10240,32,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,4096,0.027248889207839966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,3072,0.02201066745652093
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,16384,1536,0.013624889155228933
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,65536,0.35165778795878094
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,65536,0.19825155205196807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,16384,0.04924622178077698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,16384,0.10419822401470608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,12288,0.042103999190860324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,16384,0.06319022178649902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,12288,0.06722488668229845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,10240,0.056243558724721275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,10240,0.03366311060057746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,12288,0.04971111151907179
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,65536,0.20867466926574707
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,8192,0.046646220816506274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,8192,0.027936001618703205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,10240,0.03914489017592536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,7168,0.04132089018821716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,8192,0.03278399838341607
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,7168,0.02987733483314514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,6144,0.03674222363366021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,5120,0.030914667579862807
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,6144,0.026847999956872728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,5120,0.019440889358520508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,4096,0.024533333049880132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,5120,0.027099554737408955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,4096,0.016558221644825406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,3584,0.021838222940762837
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,4096,0.020277332928445604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,3584,0.018812444474962022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,3072,0.014376888672510782
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,3072,0.019277332557572257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,3072,0.017398221625222098
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,2560,0.01273155543539259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,2560,0.01662400033738878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,2560,0.016062221593326993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,2048,0.01142133358452055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,2048,0.013884444203641681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,2048,0.014334221680959066
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,1536,0.011228444675604502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,1536,0.009722666607962714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,1536,0.01276533305644989
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,1024,0.008235555556085374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,1024,0.008404444489214156
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,1024,0.010975999964608086
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,768,0.007096000015735626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,768,0.00721244431204266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,768,0.009010666774378883
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,512,0.005021333280536863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,512,0.005702222386995952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,512,0.007423111134105259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,256,0.004555555681387584
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,256,0.0034604445099830627
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,256,0.006816888848940532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,128,0.0041964443193541635
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,128,0.003050666716363695
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,64,0.0028008888992998335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,128,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,8192,32,0.0028328889360030494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,64,0.006521777974234686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,8192,32,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,65536,0.3205644554562039
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,65536,0.18075733714633516
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,12288,5120,0.025851555996470984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,16384,0.05016533202595181
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,16384,0.08285244305928548
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,65536,0.19626222716437447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,16384,0.04849688874350654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,12288,0.062084443039364286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,12288,0.03688177797529433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,12288,0.037167112032572426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,10240,0.03122044603029887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,10240,0.052592890130149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,10240,0.030953778160942927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,8192,0.025941333836979334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,8192,0.04296444522009956
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,8192,0.025847110483381484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,7168,0.023491554790072974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,7168,0.038527111212412514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,7168,0.02341777748531765
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,6144,0.033979554971059166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,6144,0.021007110675175984
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,6144,0.020948444803555805
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,5120,0.01806311143769158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,5120,0.029208001163270738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,5120,0.018550222118695576
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,4096,0.015394666128688388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,4096,0.022200889057583276
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,4096,0.015457777513398064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,3584,0.019926221834288705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,3584,0.014317333698272705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,3072,0.013351110948456658
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,3072,0.01760088900725047
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,3072,0.013175110850069257
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,2560,0.011834666960769229
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,2560,0.015232889188660515
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,2560,0.011975111232863532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,2048,0.010727110836240979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,2048,0.012667555775907306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,2048,0.010732444624106089
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,1536,0.009284444153308868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,1536,0.010448888772063784
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,1536,0.009447111023796929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,1024,0.007813333637184566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,1024,0.007859555383523306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,1024,0.008090666598743862
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,768,0.006158222340875202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,768,0.006684444430801604
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,768,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,512,0.004947555561860402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,2560,0.014381332529915703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,512,0.005392888767851724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,512,0.005469333380460739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,256,0.0033448889023727844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,256,0.004492444296677907
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,256,0.005096000101831225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,128,0.0029884444342719186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,128,0.004108444270160464
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,64,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,128,0.004879999905824661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,64,0.004910222358173794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,7168,32,0.0027706666539112725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,7168,32,0.004874666531880696
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,10240,768,0.008181333541870117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,65536,0.26372888353135854
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,16384,0.08233688937293158
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,65536,0.1519386635886298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,65536,0.15564178095923528
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,16384,0.04139111108250088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,12288,0.05510400070084465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,12288,0.03281866510709127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,12288,0.03245155678855048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,10240,0.046068443192376025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,10240,0.02782755427890354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,10240,0.028006222512986925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,8192,0.03754577702946133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,8192,0.023267555567953322
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,7168,0.03349333339267307
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,8192,0.023385778069496155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,7168,0.022361778550677832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,6144,0.029038220643997192
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,7168,0.02114222281508976
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,6144,0.019655111763212416
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,6144,0.019254222512245178
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,5120,0.01647555496957567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,5120,0.024663110574086506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,4096,0.02053955528471205
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,5120,0.016892444756295946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,4096,0.014966222974989148
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,4096,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,3584,0.018565333551830716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,3584,0.014870221416155497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,3072,0.016295111841625638
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,3584,0.01330311099688212
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,3072,0.01350488927629259
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,3072,0.01238044434123569
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,2560,0.014198222094111972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,2560,0.011952888634469775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,2560,0.011320888996124268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,2048,0.012086222569147745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,2048,0.010690666735172272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,2048,0.010081777969996134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,1536,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,1536,0.009264888862768808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,1536,0.009172444542249044
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,1024,0.007357333269384172
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,1024,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,1024,0.00674044465025266
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,768,0.006074666563007567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,768,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,768,0.006022222340106964
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,512,0.005020444591840108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,512,0.0041911109454101985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,512,0.00546577779783143
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,256,0.004564444637960858
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,256,0.0032640000184377036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,256,0.005095110999213325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,128,0.0042257776690853965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,128,0.002938666691382726
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,64,0.0027235555979940626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,128,0.004879111217127906
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,6144,32,0.0027546667390399506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,64,0.004848888764778773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,6144,32,0.004861333303981357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,6144,0.022468444373872545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,7168,0.025062221619817946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,16384,0.03937244415283203
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,65536,0.22412888209025064
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,16384,0.05685333410898844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,8192,3584,0.015801777442296345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,12288,0.04420622189839681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,16384,0.04046488801638285
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,12288,0.030028445853127375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,65536,0.13350755638546413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,10240,0.037288000186284385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,12288,0.0285537772708469
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,10240,0.026437333888477747
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,65536,0.13219111495547825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,8192,0.03046311272515191
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,10240,0.02458577851454417
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,8192,0.02069511181778378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,7168,0.027002667387326557
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,7168,0.019714666737450492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,7168,0.018886221779717337
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,6144,0.024288889434602525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,6144,0.01806400054030948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,5120,0.020630222227838304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,5120,0.015297777122921415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,6144,0.01715022159947289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,5120,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,4096,0.016977777083714802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,4096,0.01296622223324246
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,3584,0.013934221532609729
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,3584,0.015253333581818474
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,3584,0.012228444218635559
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,3072,0.013566222455766467
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,3072,0.012693333129088083
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,3072,0.01129244433508979
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,2560,0.01109155515829722
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,2560,0.011830222275522021
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,2048,0.00998488896422916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,2560,0.010435555544164447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,2048,0.00998222248421775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,2048,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,1536,0.008423111504978603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,1536,0.008707555631796518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,1536,0.008456000023418004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,1024,0.006568888823191325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,1024,0.006454222318198945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,1024,0.006079111248254776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,768,0.005715555614895291
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,768,0.005736888696750005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,512,0.0037191112836201987
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,512,0.004862222406599256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,512,0.005394666559166378
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,256,0.004462222258249919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,256,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,256,0.005085333353943295
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,128,0.002895111011134254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,128,0.0041546668443414895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,64,0.0026888888743188647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,128,0.004830222162935468
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,64,0.004845333182149463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,5120,32,0.0027146666414207886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,5120,32,0.004845333182149463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,65536,0.18349777327643502
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,65536,0.11060888899697198
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,16384,0.033949332104788885
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,16384,0.04794133371777005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,65536,0.11626488632626003
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,16384,0.030163556337356567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,12288,0.02773688899146186
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,12288,0.0381297767162323
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,12288,0.024503111839294434
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,10240,0.02275377843115065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,10240,0.03222489025857713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,8192,0.026868444350030687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,10240,0.021254221598307293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,8192,0.020115555988417733
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,8192,0.017755554782019723
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,7168,0.024082667297787134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,7168,0.020015110572179157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,7168,0.01645955608950721
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,6144,0.02162844439347585
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,6144,0.017497777938842773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,6144,0.015104888214005364
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,5120,0.015141333142916361
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,5120,0.01903377804491255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,5120,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,4096,0.015992000699043274
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,4096,0.012683555483818054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,4096,0.011739555332395764
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,3584,0.01290844463639789
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,3584,0.013605333036846586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,3584,0.011015111373530494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,3072,0.01236977759334776
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,3072,0.011585777832402123
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,3072,0.01016266644001007
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,2560,0.010938666760921478
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,2560,0.010124444133705562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,2560,0.009573332965373993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,2048,0.008932444784376357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,2048,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,2048,0.008766222331258986
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,1536,0.008191111187140146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,1536,0.007458666960398356
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,1536,0.007086222370465596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,1024,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,1024,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,1024,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,768,0.0058062221441004015
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,768,0.0046311111913787
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,768,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,512,0.004858666823969947
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,512,0.003488000068399641
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,512,0.005116444494989184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,256,0.004431999805900785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,256,0.0030862221287356485
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,256,0.004860444615284602
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,4096,128,0.004143110993835661
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,128,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,128,0.0047608887155850725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,64,0.0026782221264309357
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,64,0.0047626665068997275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,4096,32,0.0026568888376156488
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,4096,32,0.0047324444684717385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,65536,0.15443110466003418
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,7168,3584,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,65536,0.10468088918262058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,16384,0.042782220575544566
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,16384,0.031238221459918555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,12288,0.02535999980237749
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,16384,0.027772444817754958
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,65536,0.10366577572292751
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,12288,0.033320890532599554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,12288,0.0226275556617313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,10240,0.028871999846564397
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,10240,0.02124711043304867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,10240,0.019918221566412184
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,8192,0.019274666905403137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,8192,0.02404444416364034
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,8192,0.016558221644825406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,7168,0.02181600034236908
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,7168,0.015455111861228943
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,6144,0.016599999533759225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,6144,0.019341333044899833
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,6144,0.014139556222491793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,5120,0.014270222849316068
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,5120,0.017143999536832173
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,5120,0.012847111456924014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,4096,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,4096,0.012913777596420713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,4096,0.011140444212489657
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,3584,0.012538666526476542
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,3584,0.011791999969217511
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,3584,0.010557333628336588
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,3072,0.010503110786279043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,3072,0.01145866678820716
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,3072,0.0099653336736891
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,2560,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,2560,0.00923111124171151
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,2560,0.010451555252075195
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,2048,0.009093333449628618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,2048,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,2048,0.008275555239783393
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,1536,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,1536,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,1536,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,1024,0.00573244442542394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,1024,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,1024,0.005511999958091312
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,768,0.003996444245179494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,768,0.005534222142563925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,768,0.0053288886944452924
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,512,0.0034640000926123727
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,512,0.004729777574539185
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,512,0.005118222286303838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,256,0.00304888892504904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,256,0.004335999902751711
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,256,0.004837333328194088
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,128,0.0028026666906144884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,128,0.004077333129114575
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,128,0.004880889008442561
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,64,0.002630222174856398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,64,0.0047564444442590075
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3584,32,0.0026497778793176017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3584,32,0.0046986668474144405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,65536,0.13755910926394993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,6144,16384,0.040991110934151545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,65536,0.10107555654313828
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,16384,0.03819377885924445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,16384,0.02990844514634874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,16384,0.025674665967623394
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,65536,0.09143377674950494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,12288,0.023941333095232647
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,12288,0.02981688910060459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,10240,0.02601066728432973
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,12288,0.02112355497148302
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,10240,0.019882667395803664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,8192,0.021687111920780603
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,10240,0.018481777773963082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,8192,0.01830666595035129
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,7168,0.01959733333852556
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,8192,0.015791111522250705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,7168,0.017828444639841717
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,6144,0.01741333305835724
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,7168,0.014560888210932413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,6144,0.015473778049151102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,6144,0.01348977784315745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,5120,0.015431111057599386
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,5120,0.012847999731699625
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,5120,0.012143111063374413
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,4096,0.01405244403415256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,4096,0.01218933363755544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,4096,0.010611555642551847
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,3584,0.012880889077981314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,3584,0.011942221886581846
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,3584,0.010103111465771994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,3072,0.011744889120260874
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,3072,0.010948444406191507
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,3072,0.009510222408506606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,2560,0.009735110733244155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,2560,0.008765333228641087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,2560,0.008922666311264038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,2048,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,2048,0.007178666690985362
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,2048,0.008074666890833113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,1536,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,1536,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,1536,0.006495999793211619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,1024,0.004628444297446145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,1024,0.005768888940413793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,1024,0.005463111317820019
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,768,0.0038382220599386427
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,768,0.0053048887186580235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,768,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,512,0.0034711110509104202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,512,0.00463733325401942
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,512,0.005063111169470681
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,256,0.003055111194650332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,256,0.004285333471165763
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,256,0.004873777843183941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3072,128,0.004041777716742622
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,128,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,64,0.00263288886182838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,128,0.004699555536111196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,64,0.004695999953481886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,3072,32,0.0026400000270869997
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,3072,32,0.0047004446387290955
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,65536,0.11389511161380345
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,16384,0.027492443720499676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,65536,0.08016710811191134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,65536,0.07824444770812988
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,16384,0.032487998406092324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,12288,0.025942222939597234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,16384,0.02369600037733714
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,12288,0.02237422267595927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,12288,0.019384000036451552
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,10240,0.02238311039076911
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,10240,0.019153777096006606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,10240,0.017320889565679762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,8192,0.018858666221300762
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,8192,0.01696711116366916
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,8192,0.014651555154058667
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,7168,0.013838222457302941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,7168,0.01648355523745219
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,7168,0.017089777522616915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,6144,0.013769778112570444
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,6144,0.015488000379668342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,5120,0.013825777504179211
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,6144,0.012684444586435953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,5120,0.012853333519564735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,5120,0.011584888729784222
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,4096,0.012030222349696688
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,4096,0.010107555323176915
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,4096,0.010783111055692038
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,3584,0.011176000038782755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,3584,0.011238222320874533
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,3584,0.009642666412724389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,3072,0.01035377797153261
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,3072,0.010303110712104374
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,3072,0.009152000149091085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,2560,0.008023111356629265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,2560,0.009029332962301042
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,2560,0.00793333351612091
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,2048,0.0069653333889113525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,2048,0.007257777783605788
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,2048,0.0064035554726918536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,1536,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,1536,0.006186666587988536
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,1536,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,1024,0.004221333397759331
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,1024,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,1024,0.005471999860472149
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,768,0.0037600000699361167
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,768,0.005175999883148405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,768,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,512,0.003391999958289994
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,512,0.005068444543414646
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,512,0.004552888787455029
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,256,0.0042160000238153665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,256,0.0030337776988744736
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,256,0.004870222260554631
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,8192,0.02143999934196472
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,128,0.0027955555253558685
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2560,128,0.004035555654101902
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,128,0.004741333425045013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,64,0.002601777720782492
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,64,0.004691555682155821
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2560,32,0.002622222320901023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2560,32,0.0047608887155850725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,4096,0.013591110706329346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,16384,0.02810488806830512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,65536,0.09231644206576878
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,16384,0.026691555976867676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,12288,0.021572444174024794
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,16384,0.022143999735514324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,12288,0.022879999544885423
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,65536,0.0721013347307841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,65536,0.08734222253163655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,12288,0.01833599971400367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,10240,0.018799111247062683
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,10240,0.019869332512219746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,10240,0.0162124451663759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,8192,0.015969778100649517
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,8192,0.013831111292044321
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,8192,0.017638223038779367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,5120,768,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,7168,0.014500444134076437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,7168,0.01792711185084449
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,7168,0.01312355531586541
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,6144,0.013186666700575085
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,6144,0.015490666031837463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,6144,0.012108444339699216
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,5120,0.01220355596807268
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,5120,0.013608888619475894
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,4096,0.010793777803579966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,4096,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,4096,0.009840889109505547
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,5120,0.012079111403889127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,3584,0.010067555639478896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,3584,0.012432888978057437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,3584,0.008959111240175035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,3072,0.010791111323568555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,3072,0.008245333201355403
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,3072,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,2560,0.0074551113777690465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,2560,0.0068711112770769335
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,2560,0.008821333448092142
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,2048,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,2048,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,2048,0.007852444218264686
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,1536,0.005132444616821078
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,1536,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,1536,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,1024,0.003990222182538775
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,1024,0.006095110956165526
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,1024,0.005478222337034013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,768,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,768,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,512,0.0032835555159383347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,512,0.005067555440796746
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,256,0.002966222249799304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,512,0.0047208890318870544
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,256,0.004301333179076512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,256,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,128,0.0027235555979940626
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,128,0.004086222085687849
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,64,0.00258577780591117
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,64,0.0046826667255825466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,128,0.004754666652944353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,2048,32,0.0026275554878844153
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,2048,32,0.0046737777690092725
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,16384,0.021659556362364028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,16384,0.026616000466876562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,65536,0.06962133116192289
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,16384,0.02051999999417199
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,12288,0.02108000053299798
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,12288,0.017620444297790527
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,65536,0.08243378003438313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,65536,0.06625511248906453
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,10240,0.015559110376569959
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,12288,0.017027555240525138
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,10240,0.017231111725171406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,8192,0.013447110851605734
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,10240,0.015398222539159985
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,8192,0.013086222112178802
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,8192,0.0159244438012441
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,7168,0.014736889137162102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,7168,0.01259200026591619
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,7168,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,6144,0.011405333048767514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,6144,0.013654222091039022
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,6144,0.011693333586057028
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,5120,0.01055111073785358
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,5120,0.01073333372672399
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,5120,0.012295111185974546
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,4096,0.009382222261693742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,4096,0.009496889180607265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,4096,0.008635555704434713
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,3584,0.008964444200197855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,3584,0.00817599975400501
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,3584,0.010051555931568146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,3072,0.008476444416575963
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,3072,0.007248000138335758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,3072,0.00924622184700436
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,2560,0.007080000307824876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,2560,0.007800000409285228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,2560,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,2048,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,2048,0.0074417781498697065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,2048,0.0063893331421746155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,1536,0.004612444589535396
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,1536,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,1024,0.003957333250178231
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,1024,0.005534222142563925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,1024,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,768,0.003604444364706675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,768,0.005429333282841577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,768,0.005286222116814719
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,512,0.0032799999333090256
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,512,0.004507555729813046
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,512,0.005096000101831225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,256,0.002974222310715251
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,256,0.004843555390834808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,256,0.004247111164861255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,128,0.0027377777215507296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,128,0.0047280001971456735
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,128,0.0039813332259655
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,64,0.002576888849337896
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,64,0.0047271110945277745
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1536,32,0.0026062221990691293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1536,32,0.0047351109484831495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,65536,0.048650665415657886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,65536,0.06270133124457465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,65536,0.06616710954242282
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,16384,0.020433778564135235
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,16384,0.0162808895111084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,16384,0.019736889335844252
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,12288,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,12288,0.013639999760521783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,12288,0.01627200014061398
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,10240,0.012329777909649743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,10240,0.014852444330851236
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,10240,0.01481066644191742
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,8192,0.010755555497275459
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,8192,0.013192888763215808
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,8192,0.012622222304344177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,7168,0.010243555737866296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,7168,0.013208889298968844
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,7168,0.012053333222866058
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,6144,0.011252444651391772
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,6144,0.009312888814343346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,6144,0.010977778169843884
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,5120,0.008737777670224508
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,5120,0.00903555585278405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,5120,0.008906666603353288
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,4096,0.008279110822412703
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,4096,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,4096,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,3584,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,3584,0.007910221815109253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,3584,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,3072,0.007233777807818518
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,3072,0.007692444655630324
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,3072,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,2560,0.006883555402358373
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,2560,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,2560,0.006782222125265334
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,2048,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,2048,0.0063697778516345555
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,2048,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,1536,0.004603555632962121
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,1536,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,1024,0.003935111065705617
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,1024,0.005088888936572605
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,1024,0.005473777651786804
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,768,0.004886222382386525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,768,0.003587555554178026
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,768,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,512,0.0033031110134389666
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,512,0.00435644429590967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,512,0.005104889058404499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,256,0.002966222249799304
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,256,0.004102222207519743
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,256,0.004878222114510006
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,128,0.004031110968854693
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,128,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,64,0.002895999906791581
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,128,0.0046826667255825466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,64,0.004716444346639845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,1024,32,0.0028906667398081887
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,1024,32,0.004716444346639845
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,65536,0.038057777616712786
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,65536,0.06544444296095106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,16384,0.019897777173254225
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,65536,0.061441779136657715
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,16384,0.013619555367363824
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,16384,0.019486222002241347
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,12288,0.01610133383009169
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,12288,0.011777777638700275
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,12288,0.01592088904645708
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,10240,0.0144177774588267
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,10240,0.010853333605660332
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,10240,0.014558222558763293
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,8192,0.00979644474056032
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,8192,0.011714666253990598
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,7168,0.010959111154079437
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,7168,0.009038222332795462
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,7168,0.010553778045707278
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,6144,0.009757333331637913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,6144,0.008365333080291748
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,6144,0.009577777650621202
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,5120,0.00849866701496972
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,5120,0.008166222108734978
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,5120,0.008933333059151968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,4096,0.007690666450394525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,4096,0.007041778001520369
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,4096,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,3584,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,3584,0.0072266666425599
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,3584,0.007592889169851939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,3072,0.0063279999627007385
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,3072,0.007128888534175024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,3072,0.00720355576939053
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,2560,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,2560,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,2560,0.006773333168692059
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,2048,0.006096000058783426
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,2048,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,2048,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,1536,0.005392000079154968
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,1536,0.00461422238085005
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,1536,0.005859555469618903
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,1024,0.004990222139490975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,3584,7168,0.019002666076024372
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,1024,0.00396799999806616
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,1024,0.005472888963090048
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,768,0.0035777779089079965
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,768,0.0047528888616296975
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,768,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,512,0.0032968889507982465
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,512,0.004439110971159405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,512,0.005048888838953442
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,256,0.003142222141226133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,256,0.0041448887851503165
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,256,0.004807999978462855
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,128,0.003950222084919612
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,128,0.0029057777590221832
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,128,0.004709333181381226
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,64,0.002696000039577484
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,64,0.0047351109484831495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,768,32,0.002728888971938027
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,768,32,0.0047386665311124595
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,65536,0.027315555347336665
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,65536,0.06497866577572294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,65536,0.06008711126115587
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,16384,0.019722667005327012
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,16384,0.010860444770918952
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,16384,0.01862133377128177
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,12288,0.009578666753239101
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,12288,0.015749333633316886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,12288,0.015089778436554803
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,10240,0.009201777478059133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,10240,0.013553777502642738
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,10240,0.013647110925780402
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,8192,0.008451555338170793
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,8192,0.011530666715568967
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,8192,0.011576000187132092
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,7168,0.010312000082598792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,7168,0.007800888684060838
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,7168,0.010268444816271463
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,6144,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,6144,0.009536888864305284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,5120,0.006728000111050076
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,5120,0.008456000023418004
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,5120,0.008867555194430882
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,4096,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,4096,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,4096,0.007879999776681265
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,3584,0.006732444382376141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,3584,0.007165333463086023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,3584,0.007520000139872233
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,3072,0.006396444307433234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,3072,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,3072,0.007104000283612146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,2560,0.00618488879667388
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,2560,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,2560,0.006688000013430913
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,2048,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,2048,0.005387555393907759
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,2048,0.006177777631415262
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,1536,0.005352889084153705
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,1536,0.0046977777447965415
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,1536,0.0058240000572469496
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,1024,0.004907555464241239
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,1024,0.004040889028045866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,1024,0.005395555661784277
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,768,0.0036959999965296853
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,768,0.004695999953481886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,768,0.00519822206762102
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,512,0.004368000146415499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,512,0.0033688888781600525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,512,0.004980444494220945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,256,0.0030133333057165146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,256,0.004783111313978831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,128,0.003940444439649582
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,128,0.0027991111079851785
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,128,0.004648000001907349
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,64,0.0026462222966882917
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,64,0.004675555560323927
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,512,32,0.002623999905255106
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,512,32,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,65536,0.0652515557077196
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,65536,0.01688888834582435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,65536,0.05963022179073758
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,16384,0.01884355478816562
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,16384,0.008262222011884054
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,16384,0.01740533279048072
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,12288,0.015056888262430826
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,12288,0.007151111132568783
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,12288,0.014169777433077494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,10240,0.013215111361609565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,10240,0.006874666859706243
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,10240,0.012697777814335294
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,8192,0.011405333048767514
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,8192,0.006364444477690591
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,8192,0.011022222538789114
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,7168,0.010312000082598792
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,7168,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,7168,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,6144,0.009441778063774109
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,6144,0.006161777923504512
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,6144,0.009544000029563904
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,5120,0.008416888614495596
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,5120,0.006544000158707301
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,5120,0.008837333156002892
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,4096,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,4096,0.006127111199829314
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,4096,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,3584,0.007058666812049017
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,3584,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,3584,0.007507555186748505
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,3072,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,3072,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,3072,0.007161777880456712
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,2560,0.006169777777459886
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,2560,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,2560,0.006702222343948152
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,2048,0.0058551111982928375
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,2048,0.005256888767083486
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,2048,0.006178666734033161
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,1536,0.005305777821275923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,1536,0.004576888763242298
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,1536,0.005757333503829108
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,1024,0.004832889056868023
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,1024,0.0039048890272776284
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,1024,0.005356444252861871
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,768,0.004642666627963384
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,768,0.003535110917356279
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,768,0.005127999931573868
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,512,0.004334222111437055
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,512,0.003216000066863166
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,512,0.004984000076850255
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,256,0.002936888900068071
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,256,0.004791999856630961
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,128,0.003929777691761653
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,128,0.0026977778308921387
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,128,0.004659555438492033
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,64,0.0025439999169773525
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,64,0.004640000147951974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,256,32,0.0025475554996066624
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,256,32,0.004639111045334074
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,65536,0.0645128885904948
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,65536,0.011845333708657159
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,65536,0.05868533584806654
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,16384,0.018668444620238412
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,16384,0.006198222024573221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,16384,0.017231111725171406
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,12288,0.014896889527638754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,12288,0.006198222024573221
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,12288,0.014158222410413953
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,10240,0.013036444783210754
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,10240,0.006339555399285422
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,10240,0.012682666381200155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,2048,768,0.005712888720962737
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,8192,0.01129955550034841
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,8192,0.006228444476922353
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,8192,0.010986666712496014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,7168,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,7168,0.006145777801672618
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,7168,0.010310221877362993
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1536,1536,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,6144,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,6144,0.009546666509575313
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,5120,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,5120,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,5120,0.008805333740181392
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,4096,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,4096,0.007878222399287755
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,3584,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,3584,0.006330666856633292
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,3584,0.007515555454625024
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,3072,0.006511111226346757
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,3072,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,3072,0.007184888753626082
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,2560,0.006100444330109491
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,2560,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,2560,0.006748444504208035
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,2048,0.005807999935415056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,2048,0.005017777697907554
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,2048,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,1536,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,1536,0.004416000097990036
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,1536,0.005797333187527127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,1024,0.004784000002675586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,1024,0.0037324445115195382
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,1024,0.005376889059940974
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,1024,1536,0.005694222119119432
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,768,0.0035626664757728577
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,768,0.004547555413511065
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,768,0.005184000151024925
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,512,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,512,0.003256888853179084
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,512,0.004935111022657818
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,256,0.004000888930426704
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,256,0.002887111157178879
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,256,0.0047217777205838096
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,128,0.0038311113086011675
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,128,0.00272088891102208
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,64,0.002567111204067866
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,128,0.004638222356637319
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,64,0.004600000050332812
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,128,32,0.002571555475393931
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,128,32,0.004600888739029567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,65536,0.009523555636405945
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,16384,0.006044444524579578
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,65536,0.056095110045539014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,12288,0.006010666489601135
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,16384,0.017210667332013447
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,12288,0.014177777700954013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,10240,0.006206222292449739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,10240,0.012717333104875354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,8192,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,7168,0.005980444451173146
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,8192,0.011015111373530494
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,6144,0.0058711109062035876
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,7168,0.010272000398900773
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,6144,0.009506666825877296
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,5120,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,4096,0.005864000154866113
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,5120,0.008845333423879411
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,4096,0.007870222131411234
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,3584,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,3072,0.006020444548792309
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,3584,0.007568000091446771
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,2560,0.005519999812046687
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,3072,0.00721599989467197
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,2048,0.004952888935804367
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,2560,0.0067528887755341
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,2048,0.006206222292449739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,1536,0.004318222403526306
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,1536,0.005797333187527127
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,1024,0.0037164443896876443
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,1024,0.005369777894682354
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,768,0.003400000019205941
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,768,0.005150222116046482
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,512,0.003128000017669466
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,512,0.004923555586073133
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,256,0.0028613333900769553
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,256,0.0046862223082118565
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,128,0.0027262222849660446
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,128,0.00462755560874939
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,64,0.002566222308410539
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,64,0.0046302220887608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,64,32,0.0025617778301239014
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,64,32,0.0046302220887608
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,65536,0.008282666405042013
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,16384,0.00620177760720253
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,65536,0.05504088931613498
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,12288,0.006074666563007567
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,16384,0.017268444101015728
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,12288,0.014152889450391134
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,10240,0.006206222292449739
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,10240,0.012682666381200155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,8192,0.006110222389300664
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,8192,0.011001778145631155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,7168,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,7168,0.010284444524182213
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,6144,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,6144,0.009564444422721863
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,5120,0.006213333457708359
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,5120,0.00888000014755461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,4096,0.0058204444746176405
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,4096,0.007903999752468532
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,3584,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,3584,0.007520889242490132
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,3072,0.005986666513813867
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,3072,0.007204444044166141
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,2560,0.005506666584147348
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,768,8192,0.012208888928095499
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,2560,0.006779555645253923
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,2048,0.006187555690606435
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,2048,0.004938666605287128
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,1536,0.004332444320122401
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,1536,0.005843555761708154
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,1024,0.00370399985048506
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,1024,0.0053635554181204895
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,768,0.0034097778714365433
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,768,0.00518488883972168
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,512,0.0031395554542541504
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,512,0.004948444250557157
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,256,0.0028071111689011254
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,128,0.0026213334252436957
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,256,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,64,0.002542222125662698
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,128,0.004616888860861461
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,64,0.004622222234805425
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,float16,1,32,32,0.002550222186578645
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8,1,32,32,0.004628444297446145
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,6144,0.009492444495360056
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,512,256,0.004061333421203825
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,256,256,0.004059555629889171
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,6144,0.00942311104800966
SGLang,0.5.8.post1,NVIDIA H100 80GB HBM3,gemm,sglang,fp8_block,1,128,4096,0.0075786668393347
