framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,3584,0.01645866698688931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,768,0.0036195553839206696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,256,0.006041777630647023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,4096,0.019839111301634047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,10240,0.09441688987943862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,2048,0.0251217782497406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,64,0.0069120000633928515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,4096,0.007924444145626491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,3072,0.03254755669169956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,5120,0.018917333748605516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,5120,0.009668444593747457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,64,0.002508444504605399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,6144,0.04332800043953789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,256,0.00655644428398874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,5120,0.01516888870133294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,768,0.005664000080691443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,6144,0.02032088902261522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,12288,0.009834667046864828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,4096,0.008817777865462834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,1024,0.00619911112719112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,3584,0.006818666640255187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,16384,1.9245796203613281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,4096,0.07550666729609172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,7168,0.013071999781661563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,1536,0.12632532914479574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,2560,0.01553244392077128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,8192,0.03299466768900553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,8192,0.023040889037979975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,512,0.004561777744028303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,10240,0.019186667270130582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,8192,0.01239111108912362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,16384,0.021378666162490845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,512,0.004439110971159405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,1536,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,3584,0.00943022221326828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,768,0.008280889027648503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,32,0.1824124389224582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,512,0.005098666581842634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,3072,0.010598222414652506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,6144,0.02296000056796604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,3584,0.007320000065697565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,8192,0.30793510542975533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,3584,0.03176888823509216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,4096,0.009935999910036722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,7168,0.052631109952926636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,2048,0.006693333387374878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,3584,0.00777688870827357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,6144,0.008001777860853408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,12288,0.007070221834712558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,6144,0.552677313486735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,64,0.0028986665937635633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,3584,0.12521510654025608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,12288,0.007256000406212277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,65536,0.061625778675079346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,256,0.012400888734393649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,256,0.003272888975010978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,8192,0.013761777844693927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,32,0.005935110979610019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,1024,0.02992977698644002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,2048,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,7168,0.024736000431908503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,512,0.024512888656722173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,16384,0.0553288890255822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,32,0.0074240002367231585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,1024,0.03538044293721517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,3584,0.02221600049071842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,4096,0.00812711069981257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,65536,0.0743724438879225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,256,0.006584888945023219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,7168,0.010545777777830759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,256,0.010261333651012844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,4096,0.02773688899146186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,2560,0.010135110881593492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,8192,0.04316888915167915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,2048,0.006063999815119638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,4096,0.08769777748319839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,2560,0.4849866761101617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,768,0.012485333614879184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,10240,0.025607110725508794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,2560,0.10225244363149007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,512,0.0058471109304163195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,10240,0.06768088870578341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,512,0.007755555212497711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,5120,0.02239111065864563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,1024,0.006497777584526274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,128,0.005072888814740711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,128,0.004663111021121343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,16384,0.018741332822375827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,4096,0.012042666474978128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,256,0.0032684444967243406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,3072,0.019468444916937087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,128,0.006254222244024277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,6144,0.01312622262371911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,32,0.0034560000316964257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,65536,0.03815999958250258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,6144,0.04695377747217814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,2048,0.09921333524915908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,65536,0.05736800034840902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,4096,0.024479111035664875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,5120,0.009291555318567488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,128,0.002761777697337998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,3072,0.007161777880456712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,4096,0.019631110959582858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,6144,0.012490666574902005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,16384,0.29078488879733616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,8192,0.030517333083682593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,1536,0.0063546668324205614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,64,0.0026364444444576898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,3584,0.009744889206356471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,3072,0.00831200016869439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,64,0.006236444330877728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,1024,0.012074666718641916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,512,0.011319110790888468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,768,0.006034666465388403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,32,0.002761777697337998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,3072,0.060871111022101514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,3072,0.030481778913074072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,1024,0.005631999837027655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,512,0.004403555558787452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,32,0.0028533333291610083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,2048,0.005361777626805835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,3072,2.5822132958306208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,128,0.009666666388511658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,256,0.03137066629197862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,65536,0.8972400029500326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,65536,0.6058728959825304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,512,0.00795377790927887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,2048,0.008805333740181392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,1024,0.010270222193664974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,5120,0.11135110590193008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,1536,0.010545777777830759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,2048,0.01571911076704661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,8192,0.05158488949139913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,6144,0.01200177768866221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,8192,0.14969421757592097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,256,0.004458666675620609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,32,0.006311111152172089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,7168,0.01054133309258355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,5120,0.016561778055297006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,32,0.04606222112973531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,512,0.01365600029627482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,8192,0.03163466519779629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,2560,0.005884444548024072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,16384,0.4991520245869954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,256,0.022563555174403723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,6144,0.06518844763437907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,128,0.0061039999127388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,128,0.005416888743638992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,2048,0.006829333388143116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,7168,0.013742222554153867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,256,0.003156444471743372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,256,0.003887999802827835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,7168,0.10789955324596828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,64,0.004981333182917701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,768,0.005021333280536863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,3072,0.016701334052615695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,10240,0.015658666690190632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,8192,0.5349315537346734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,1024,0.20086044735378691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,10240,0.03933066791958279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,3584,0.0064479998416370815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,3584,0.007277333074145847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,2560,0.0233297778500451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,128,0.004044444610675176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,3584,0.008570666942331526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,12288,0.013112000293201871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,128,0.0035777779089079965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,3072,0.01513066722287072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,5120,0.37203200658162433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,7168,0.015840000576443143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,2048,0.006766222417354584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,2560,0.01548266741964552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,3584,0.008298666940795051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,65536,0.36026843388875324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,2560,0.012107555237081317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,65536,0.07618311378690931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,8192,0.06274844540490045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,1536,0.0162426663769616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,256,0.00831733312871721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,3584,0.013910222384664746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,256,0.003145777723855443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,128,0.005433777968088786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,512,0.003400000019205941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,16384,0.006624889042642381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,4096,0.017196445001496207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,2048,0.007300444775157505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,32,0.005543111099137201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,1536,0.006730666591061487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,5120,0.009688888986905416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,3072,0.22375644577874076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,1024,0.017813333206706576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,2048,0.011490667031870948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,2560,0.006893333461549547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,2048,0.02858755654758877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,1536,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,64,0.0027919999427265594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,3584,0.007882666256692674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,10240,0.35034754541185165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,256,0.00573511090543535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,768,0.16003644466400146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,768,0.006211555666393704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,65536,0.17972800466749403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,8192,0.010224888722101847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,128,0.005384888913896348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,10240,0.03596977723969354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,2048,0.006653333289755716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,1536,0.01163733336660597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,6144,0.025825776987605627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,4096,0.006480888773997624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,1024,0.0087333329849773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,5120,0.01127199994193183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,7168,0.035298665364583336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,256,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,10240,0.312096012963189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,1536,0.031091554297341242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,3584,0.031523555517196655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,3584,0.010235555469989777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,12288,0.012459555433856117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,8192,0.03487733337614272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,5120,0.011779555843936073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,3072,0.0107004443804423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,16384,0.007673777639865875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,16384,0.3507644335428874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,6144,0.051051557064056396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,512,0.0075733330514695905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,6144,0.1316924492518107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,1536,0.006727111008432176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,1024,0.010326222413116032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,1536,0.008256889051861232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,3072,0.010843555960390302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,2560,0.32374755541483563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,256,0.005292444593376583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,4096,0.014535999960369535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,8192,0.015642666154437594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,2560,0.007600000335110559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,5120,0.01516266663869222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,12288,0.11448444260491265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,3072,0.010346666806273991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,16384,0.02296000056796604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,4096,0.006172444257471297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,1024,0.003968888686762916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,3584,0.014955555399258932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,10240,17.002682156032986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,512,0.004387555436955558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,128,0.003909333298603694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,16384,0.14455644289652506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,2048,0.02943555514017741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,6144,0.007618666523032718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,3584,0.008336000144481659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,32,0.0068711112770769335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,6144,0.022054221895005967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,2048,0.008553778131802877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,3584,0.025713778204388086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,2560,0.006743999818960826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,32,0.006626666833957036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,16384,0.018411555223994784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,3072,0.009813333551088968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,12288,0.005941333456171884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,4096,0.012884444660610624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,32,0.002623999905255106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,5120,0.15947288937038845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,1536,0.008398222426573435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,512,0.005684444473849402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,7168,0.023974221613672044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,3072,0.047855999734666615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,16384,0.02751111156410641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,16384,0.025255999631351892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,65536,0.3857368893093533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,16384,0.006031999985376994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,32,0.005170666509204441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,6144,0.8771235677931043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,1536,0.007717333734035492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,32,0.009522666533788046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,768,0.013774221969975365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,512,0.0041777777175108595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,10240,0.02072622213098738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,64,0.0030177777840031516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,10240,0.006483555667930179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,65536,0.8064853350321451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,65536,1.2304346296522353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,1536,0.00608711110221015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,768,0.011844444606039258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,2560,0.0081377774477005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,32,0.00351200004418691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,768,0.00552977787123786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,512,0.0032559999575217566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,32,0.002733333243264092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,768,0.008718222379684448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,1536,0.05156355433993869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,5120,0.00871288859181934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,2048,0.0052622221410274506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,7168,0.03862844573126899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,512,0.005314666777849197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,12288,0.01735377808411916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,8192,0.01204711116022534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,3584,0.14120622475941977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,7168,0.07450133562088013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,3584,0.008988444175985124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,5120,0.01242844429281023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,10240,0.13672622044881186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,2560,0.019730667273203533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,8192,0.0617777771419949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,768,0.003589333345492681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,4096,0.3141546779208713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,256,0.005473777651786804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,3072,0.027748444014125403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,65536,0.353948434193929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,10240,0.03751555416319106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,10240,0.07766577932569715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,16384,0.018785778019163344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,2560,0.013226666384273104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,8192,0.015270221564504834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,16384,0.06070311201943291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,128,0.013799111048380533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,16384,0.029693331983354356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,2048,0.012230222423871359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,3072,0.1377991173002455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,32,0.0032871110985676446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,2560,0.007155555817815993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,512,0.0069360000391801195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,3072,0.1372577746709188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,1536,0.009883555273214975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,512,0.01019200020366245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,65536,4.027529822455512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,3584,0.007096000015735626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,5120,0.03407377666897244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,7168,0.00657155571712388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,1536,0.0550799998972151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,1024,0.020793777373101976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,10240,0.027282666828897264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,5120,0.219833771387736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,64,0.005128889034191768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,128,0.004058666527271271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,512,0.0037324445115195382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,3072,0.015589333242840238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,8192,0.047800001170900136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,2048,0.007693332930405934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,6144,0.014222222897741528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,7168,0.02527644402450985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,8192,0.01420088940196567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,3584,0.010799999866220685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,5120,0.018096889058748882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,8192,0.14206400182512072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,768,0.00553688903649648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,12288,0.04839644498295254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,512,0.00832088871134652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,3584,0.006715555571847492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,768,0.005128889034191768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,6144,0.027422222826215956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,768,0.008788444101810455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,1536,0.11908977561526829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,1536,0.03603822324011061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,2048,0.27635733286539715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,8192,0.055030223396089345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,10240,0.017973333597183228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,10240,0.016656888855828177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,2560,0.014577777849303352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,8192,0.08639111121495564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,6144,0.01939288940694597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,128,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,65536,0.18800977865854898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,1536,0.016872000363137987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,65536,0.1167715522978041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,1536,0.004635555462704765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,1536,0.005433777968088786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,16384,0.044884443283081055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,128,0.05390310949749417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,512,0.0034346667428811393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,128,0.006061333335108227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,6144,0.01035733355416192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,7168,0.035903112755881414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,4096,0.011758222348160215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,32,0.006225777582989798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,12288,0.062281780772738986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,12288,0.039078222380744085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,6144,0.007689778175618913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,1536,0.021689777572949726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,256,0.011633777783976661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,3584,0.035197334157096014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,10240,0.03387022349569533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,7168,0.20856355296240914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,16384,0.06956622335645887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,1024,0.04233155647913615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,7168,0.026672000686327618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,2048,0.04725777771737841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,65536,0.26265332433912486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,10240,2.10317251417372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,3072,0.014331556028789945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,768,0.009892444643709395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,64,0.0032266666077905228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,2560,0.007017778025733099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,2048,0.01587288909488254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,3584,0.007098666495747036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,32,0.005285333428117964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,4096,0.17114844587114122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,12288,0.01535377734237247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,4096,0.03308533297644721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,2048,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,512,0.020821332931518555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,3584,0.010850666297806634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,1536,0.02379911144574483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,768,0.006064888917737537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,7168,0.029696000946892634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,6144,0.02816533380084568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,32,0.005886222339338726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,6144,0.05282844437493218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,64,0.056516442033979625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,4096,0.009498666558000777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,2048,0.03821511069933573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,512,0.03186488813824124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,32,0.008455110920800103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,64,0.005079999979999331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,1024,0.01755644381046295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,3072,0.008524444368150499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,16384,0.05505422088834974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,1536,0.029095109966066148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,4096,0.041307555304633245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,16384,0.021389333738221064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,3072,0.0445128877957662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,3584,0.014039110806253223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,7168,0.0557840002907647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,8192,0.03436355458365546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,2048,0.06584355566236708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,5120,0.01129333343770769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,16384,0.05226933293872409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,10240,0.014999111493428549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,6144,0.017902221944597032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,3584,0.039080000585979886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,128,0.0041999999019834726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,1536,0.005391110976537068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,5120,0.011681777735551199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,2048,0.006686222222116258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,7168,0.013834666874673633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,3072,0.009151111046473185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,1536,0.004331555631425646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,4096,0.0058373332851462895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,256,0.028736889362335205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,2048,0.01313866674900055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,8192,0.2185493310292562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,256,0.005113777601056629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,32,0.008201777935028076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,256,0.004368888835112254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,2048,0.006704888823959563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,256,0.004292444636424382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,768,0.004110222061475118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,8192,0.049076444572872586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,7168,0.024904888537194993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,65536,0.06949866480297513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,10240,0.018713777263959248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,16384,0.01864622202184465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,512,0.004662222332424588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,2560,0.024015110399987962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,10240,0.013645333548386892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,768,0.014925332532988654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,128,0.006135111053784688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,4096,0.008479110896587372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,10240,0.0751004417737325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,12288,0.03112444281578064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,128,0.02124622298611535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,2048,0.01277511070171992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,3584,0.018340445227093168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,64,0.006308444258239534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,128,0.005199111170238919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,65536,0.16602222124735513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,4096,0.09245600303014119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,4096,0.00961511085430781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,65536,0.05406666795412699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,4096,0.009970666633711921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,32,0.024392889605628118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,512,0.003547555456558863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,2048,0.0070053330726093715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,64,0.003403555601835251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,256,0.004948444250557157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,8192,0.03157155712445577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,5120,0.01457333399189843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,256,0.13296533955468073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,512,0.003440888805521859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,8192,0.029729776912265356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,10240,0.009375111096435124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,5120,0.008809777597586313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,768,0.005482666608360078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,16384,0.00609688874748018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,12288,0.5297662417093912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,3072,0.016082665986484952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,64,0.002529777793420686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,3584,0.006830222076839871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,256,0.009373333719041612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,10240,0.22011378076341417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,768,0.007724444071451823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,2560,0.016403555870056152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,256,0.004569777597983678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,65536,0.48108710183037656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,8192,0.015205333630243937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,8192,0.06315022044711642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,65536,0.07691466808319092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,128,0.004895110925038655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,6144,0.1553973356882731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,1024,0.006335999816656113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,64,0.004981333182917701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,10240,0.016770665844281513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,6144,0.032002667586008705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,1536,0.006183111005359226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,7168,0.01054133309258355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,10240,0.0350471105840471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,5120,0.01370666672786077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,3584,0.00832444429397583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,6144,0.11228888564639622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,3584,0.013236444857385425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,6144,5.221719106038411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,128,0.004264889078007804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,7168,0.05493599838680691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,16384,0.036460445986853704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,512,0.003682666768630346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,3072,0.010763555765151978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,256,0.002983999955985281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,64,0.010401777923107147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,2560,0.05387466814782885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,4096,0.03543555405404832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,128,0.00304888892504904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,4096,0.04414933257632785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,5120,0.006812444577614467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,1536,0.015276443627145557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,768,0.010176888770527309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,2560,0.022368000613318548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,3072,0.021704889006084863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,7168,0.061703999837239586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,64,0.002557333351837264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,512,0.005461333526505365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,65536,0.06587022542953491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,256,0.006248000181383557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,5120,0.01146488885084788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,64,0.005452444569932089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,512,0.010878221856223213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,512,0.006002666635645761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,32,0.0025111111915773815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,1536,0.008650666309727563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,10240,0.007286222444640265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,12288,0.017858665850427415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,2048,0.010178666975763109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,4096,0.027127110295825537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,4096,0.014849777022997538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,3072,0.0075333333677715724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,256,0.003990222182538775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,3584,0.006931555353932911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,32,0.0026337777574857077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,6144,0.03822755482461717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,512,0.005521777603361342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,128,0.005196444276306364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,1024,0.004649777793222004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,3584,0.009672000176376766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,12288,0.06460177898406982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,768,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,1536,0.04208711120817396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,65536,0.0711759991115994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,2560,0.08911910984251235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,1536,0.006743999818960826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,16384,0.00963288876745436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,512,0.005443555613358815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,16384,0.006021333237489064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,6144,0.014241778188281588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,2560,0.02993866801261902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,1024,0.004576000074545543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,6144,0.010829333629873065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,4096,0.008722667064931657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,1536,0.011867555479208628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,65536,0.12493067317538792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,2048,0.009184000392754873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,2560,0.012323555847009024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,8192,0.009759999811649323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,65536,0.2019395563337538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,2048,0.01070844464831882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,6144,0.022672888305452134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,2560,0.019904888338512845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,1536,0.006450666735569636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,12288,0.01644533375898997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,128,0.003240888938307762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,65536,0.07506844732496473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,32,0.00282577777074443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,2560,0.015140444040298462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,3072,0.011526222030321756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,1536,0.12501688798268637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,32,0.002628444383541743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,5120,0.39631909794277614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,10240,0.017157332764731515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,2048,0.008392000363932716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,3584,0.027689778142505225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,3584,0.01183199965291553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,1536,0.01440088947614034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,128,0.005913777897755305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,256,0.004400889078776042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,512,0.005552000055710475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,4096,0.014715555641386243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,8192,0.20251554912990996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,2560,0.006518222391605377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,65536,0.554991086324056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,65536,0.3389208846622043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,2048,0.010805333654085795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,768,0.005746666755941179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,1024,0.14819467067718506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,5120,0.010311110979980892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,16384,0.037532445457246565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,1536,0.006603555546866522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,1536,0.005679111099905438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,768,0.008202667037645975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,8192,0.023553777072164748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,2048,0.11156888802846272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,256,0.006507555643717448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,3584,0.006208000083764394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,32,0.005728000154097875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,1024,0.02513244416978624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,7168,0.13134311305152044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,2560,0.015351111690203348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,2560,0.012583110895421771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,5120,0.011406222151385413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,8192,0.45014487372504336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,65536,0.30784532758924693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,4096,0.01888533267709944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,1536,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,1024,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,256,0.005596444424655702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,7168,0.1353751156065199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,512,0.004843555390834808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,8192,0.027116444375779893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,10240,0.013470222552617392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,4096,0.0424924459722307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,65536,0.3141111003028022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,2560,0.008752889103359645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,5120,0.022844443718592327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,6144,0.022203556365436975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,7168,0.010777778095669217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,2048,0.03478755553563436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,32,0.0026568888376156488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,2560,0.00608711110221015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,10240,0.0798799991607666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,8192,0.017140444782045152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,1024,0.006305777778228124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,10240,0.008482666479216682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,10240,0.018944889307022095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,8192,0.017314665847354464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,16384,0.13716710938347712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,3072,0.025243555506070454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,3072,0.038337777058283486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,65536,0.07181333170996772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,64,0.027100443840026855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,3072,0.006440889090299606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,3072,0.01313866674900055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,1536,0.010684444672531553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,256,0.010690666735172272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,512,0.006128888991143968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,1024,0.009911110831631554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,7168,0.12355021635691325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,6144,0.026387555731667414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,256,0.00639111093348927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,4096,0.007889777421951294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,1024,0.00905777762333552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,32,0.0058106668293476105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,2048,0.0058311112225055695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,65536,0.06264533599217732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,65536,1.6560711330837672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,512,0.004885333279768626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,768,0.004568888909286923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,65536,0.07586222224765353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,4096,0.019494222270117868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,3584,0.043839110268486865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,5120,0.009276444713274637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,65536,0.14479645093282065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,2560,0.013257777525318993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,65536,0.4111466672685411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,768,0.015814221567577787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,32,0.019646222392718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,3584,0.020505777663654752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,2048,0.011048888994587792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,512,0.005577777822812398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,12288,0.010421333213647207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,256,0.006937777830494775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,8192,0.06335110796822442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,64,0.007485333416197035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,8192,0.34787734349568683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,6144,0.033386667569478355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,10240,0.07962133487065633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,8192,0.008628444539176093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,64,0.006443555570311016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,12288,0.03532088796297709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,6144,1.1805031034681532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,65536,0.0687555538283454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,512,0.004683555414279302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,2560,0.018953777021831937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,2560,0.019398222366968792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,512,0.008271111382378472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,3072,0.13941421773698595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,256,0.003156444471743372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,7168,0.04458399944835239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,65536,0.08186222447289361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,512,0.013271999855836233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,5120,0.01427288850148519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,4096,0.024696888195143804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,16384,0.022994667291641235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,12288,0.04365866714053684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,12288,0.02091644373204973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,4096,0.024320888850424025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,4096,0.013839999834696451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,65536,0.1902844508488973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,1536,0.005509333478079901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,256,0.004436444491147995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,10240,0.01927911076280806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,64,0.0032586666444937387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,1024,0.008868444297048781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,65536,0.060099555386437305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,2560,0.021219554874632094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,3584,0.014816888504558139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,5120,0.014631110760900708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,8192,0.05320800013012356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,7168,0.0558168888092041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,16384,0.02952622373898824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,128,0.014656888114081489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,3584,0.022026666336589392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,6144,0.020946666598320007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,1024,0.016117332710160148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,128,0.005525333185990651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,512,0.005504888792832692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,6144,0.014690667390823364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,7168,0.006490666833188798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,1536,0.036497778362698026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,512,0.005324444423119227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,64,0.005070222334729301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,3584,0.007892444729804993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,3072,0.017886221408843994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,4096,0.047578665945265025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,65536,0.4090586768256293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,5120,0.01683822274208069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,32,0.005081777771313985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,1024,0.030419554975297716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,16384,0.08908533387713963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,16384,0.039488888449139066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,32,0.0027457777824666765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,3072,0.010480889015727572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,768,0.005816000203291575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,8192,0.030115554730097454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,1024,0.003929777691761653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,2560,0.00757244477669398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,7168,0.03592711024814182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,2560,0.008202667037645975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,2560,0.011554666691356234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,8192,0.016031111280123394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,512,0.007348444726732042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,6144,0.006235555642180973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,5120,0.22536532084147134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,512,0.006093333164850871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,12288,0.040602667464150324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,512,0.004383999854326248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,3072,0.1468435525894165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,1536,0.058244446913401283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,1536,0.026260443859630164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,256,0.005149333427349727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,65536,0.06782311201095581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,12288,0.013736888766288757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,256,0.010499555203649731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,5120,0.12756533092922634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,3584,0.04758488800790575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,3072,0.013912889692518445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,3072,0.011761777930789523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,64,0.004592888885074192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,3584,0.01070577816830741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,10240,0.09275822507010566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,16384,0.01737155516942342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,8192,0.011776888536082374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,3584,0.016664889123704698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,1024,0.008168000313970778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,12288,0.030236442883809406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,512,0.0053022222386466126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,768,0.01754133403301239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,7168,0.014162666267818876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,10240,0.09675555759006077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,64,0.006521777974234686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,128,0.004628444297446145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,512,0.1415262222290039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,2560,0.014659555421935188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,7168,0.012122666670216454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,4096,0.014144889182514615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,2048,0.07204977671305339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,6144,0.047825776868396334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,6144,0.01583377851380242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,2048,0.005807999935415056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,8192,0.025117332736651104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,2048,0.007746666669845581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,6144,0.2341609001159668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,256,0.006937777830494775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,128,0.006825777805513806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,1536,0.006717333363162146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,2560,0.008739555875460306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,10240,0.05499200026194254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,2048,0.006940444310506185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,2560,0.016992888516849942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,16384,0.03463822272088792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,2048,0.008568888737095727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,6144,0.03472977876663208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,768,0.01533600025706821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,7168,0.12732356124454075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,12288,0.029709332519107397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,7168,0.017895999881956313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,32,0.0029048888633648553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,256,0.0034826666944556763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,768,0.01072088877360026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,8192,0.2541875574323866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,256,0.004609777695602841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,1536,0.005767111149099138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,768,0.012311110893885294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,3584,0.006704000135262807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,3072,0.012643555800120035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,2048,0.006022222340106964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,5120,0.04016710983382331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,5120,0.06406844324535794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,1024,0.008748444418112436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,3072,0.011567110816637674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,8192,0.152511994043986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,1536,0.0058604445722368025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,10240,0.8947298261854383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,6144,0.036658667855792575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,512,0.005132444616821078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,8192,0.08618933624691433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,512,0.005752000129885144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,128,0.0026515554636716843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,2048,0.00685066646999783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,16384,0.04988622334268358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,3072,0.010008888940016428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,128,0.002941333378354708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,7168,0.010804444551467896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,768,0.007915555602974361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,10240,0.04131733377774557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,8192,0.014114666316244336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,7168,0.02996622191535102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,7168,0.02718399961789449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,10240,0.009752888646390703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,1536,0.010823110739390055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,6144,0.011129777464601727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,16384,0.03401422169473436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,4096,0.009145778086450364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,512,0.005470222069157495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,768,0.018304889400800068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,8192,0.014359111587206522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,2560,0.011909333368142446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,256,0.004974222017659081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,128,0.019680000013775296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,16384,0.02476088868247138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,12288,0.0303457776705424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,3584,0.01440355512830946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,6144,0.020852444900406733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,1536,0.005512889060709212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,1536,0.030329777134789362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,128,0.005229333208666907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,5120,0.012929778132173749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,65536,0.3984079890780979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,1536,0.0047333331571684945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,32,0.003120888852410846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,6144,0.15882844395107693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,12288,0.029492444462246362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,512,0.0032435556252797446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,2048,0.009749333063761393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,4096,0.06980266835954455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,1536,0.007277333074145847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,1536,0.006863111009200414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,768,0.008815111385451423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,3072,0.023836443821589153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,16384,0.054422222905688815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,12288,0.02583911187118954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,512,0.006853333363930385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,3584,0.020181333025296528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,3584,0.011504000259770287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,65536,0.14050755235883924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,2560,0.009703999592198266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,3584,0.12169688277774388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,1024,0.00741244438621733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,8192,0.049997333023283214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,768,0.2231057749854194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,65536,0.1976097822189331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,4096,0.5513075722588433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,1024,0.0380951099925571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,2048,0.01258133351802826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,5120,0.02069511181778378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,1536,0.01144266708029641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,128,0.01977599991692437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,128,0.0029173334025674393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,768,0.009975999593734741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,3072,0.053222222460640796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,3072,0.009340444372759925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,512,0.0047457776963710785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,3072,0.008454221818182204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,1024,0.004659555438492033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,6144,0.010643555886215635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,6144,0.0059075554211934405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,5120,0.010710222025712332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,768,0.010401777923107147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,3584,0.01330933305952284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,2560,0.01034400032626258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,2560,0.009577777650621202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,256,0.0035377778112888336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,7168,0.03905511233541701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,1536,0.03736977775891622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,65536,0.01755288905567593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,6144,0.01349422252840466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,128,0.0033608888172441055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,12288,0.012097777591811286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,4096,0.01886133352915446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,10240,0.03359377715322707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,512,0.007240888973077138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,3072,0.01016177816523446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,10240,0.041690667470296226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,12288,0.01789777808719211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,768,0.017251556118329365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,2048,0.018956444329685636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,256,0.004713777866628435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,2048,0.09570222430759007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,16384,0.0236817780468199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,10240,0.01624000072479248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,3584,0.008782222039169734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,65536,0.018872888551818002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,4096,0.011643555429246692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,64,0.006116444451941385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,64,0.006025777922736273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,512,0.016155555844306946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,2560,0.1612106694115533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,7168,0.00602311102880372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,7168,0.03125155634350247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,4096,0.012487110992272695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,256,0.004256000121434529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,3584,0.00775022225247489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,7168,0.038776000340779625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,16384,0.06493422057893541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,256,0.009891555541091496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,7168,0.02019644445843167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,6144,0.009690666364298927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,768,0.0053324442770746015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,7168,0.006591111007663939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,10240,0.03719644414054023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,128,0.003129777808984121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,1536,0.006030222194062338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,65536,0.03164711263444688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,768,0.005668444352017508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,256,0.004310222135649787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,4096,0.02314044369591607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,2560,0.008647110727098253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,64,0.002479111154874166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,4096,0.014496000276671516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,12288,0.5744053522745768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,2048,0.009287999735938178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,64,0.0063440000845326324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,256,0.004468444320890638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,12288,0.06482844220267402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,12288,5.116415235731337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,65536,0.07947733004887898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,64,0.0064488889442549805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,4096,1.0935404035780165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,512,0.0040995557275083335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,768,0.009910222556855943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,512,0.0069644442862934535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,256,0.006806222101052602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,128,0.0032684444967243406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,7168,0.030023998684353296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,1024,0.016495111915800307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,3072,0.015246222416559855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,8192,0.06356088982688056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,768,0.03012088934580485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,4096,0.03320711188846164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,65536,0.3730657895406087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,768,0.008628444539176093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,10240,0.022999111149046157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,3072,0.012664889295895895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,3072,0.008547555241319869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,7168,0.0214684440029992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,6144,0.01627288924323188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,1024,0.005187555733654234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,768,0.005586666779385672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,3072,0.013595555391576556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,8192,0.013796444568369122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,512,0.005730666634109285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,7168,0.2250995635986328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,128,0.010100444157918295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,2048,0.0709644423590766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,12288,0.04490400022930569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,1024,0.004967111266321606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,256,0.004431999805900785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,768,0.05283733208974203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,7168,0.03759733173582289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,1536,0.006188444379303191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,4096,0.10088000032636855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,3584,0.4010675483279758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,12288,0.021077333225144282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,2560,0.016561778055297006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,16384,0.023919110496838886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,2560,0.015039111177126566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,6144,0.11068177885479397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,2560,0.007544889218277401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,4096,0.012727110750145383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,8192,0.012548444171746572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,10240,0.030360887447992962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,5120,0.011891555454995898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,5120,0.09616355763541327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,65536,0.19908800390031603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,2560,0.23625599013434517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,10240,0.015090665883488126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,6144,0.03906755646069845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,1536,0.05748266643948025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,32,0.00527555536892679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,1536,0.03352800011634827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,4096,0.018038221531444125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,8192,0.016368889146380954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,3072,0.008344000412358178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,3072,0.007450666692521837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,2560,0.11874932712978786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,65536,0.5561982260810004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,128,0.00489688871635331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,65536,0.14733510547214085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,12288,0.04144710964626736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,5120,0.02225422196918064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,12288,0.06254488892025419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,768,0.004784000002675586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,512,0.005958222266700532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,65536,0.08750044637256199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,2560,0.01401600076092614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,10240,0.031567109955681696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,1536,0.009325332939624786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,512,0.004911111046870549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,12288,0.24375287691752115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,5120,0.009003555609120263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,65536,0.11728889412350124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,1536,0.006870222174459034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,3072,0.009881777895821465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,16384,0.037647111548317805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,32,0.005126222140259213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,3584,0.007958222594526079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,32,0.005000000198682149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,7168,0.014532445205582513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,2048,0.006160000132189856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,3584,0.007898666792445714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,1536,0.008475555313958062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,12288,0.030838221311569214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,5120,0.14425422085656062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,3584,0.05504622062047323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,32,0.005019555489222209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,2560,0.012507555385430654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,256,0.004943999979231092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,6144,0.10664621988932292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,6144,0.009553777674833933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,1024,0.009286222358544668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,512,0.005454222361246745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,65536,0.014340443743599786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,5120,0.02091111077202691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,8192,0.014191110928853353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,3072,0.008007110820876228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,3584,0.014704000618722705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,4096,0.008622222476535374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,128,0.002692444456948174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,7168,0.05379288726382785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,1536,0.005444444302055571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,1536,0.018927999668651156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,2560,0.09853955772187974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,3584,0.026748443643252056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,7168,0.027971555789311726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,16384,0.012486221889654795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,256,0.003039111072818438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,1024,0.006089777582221561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,32,0.012154666913880242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,512,0.008311111066076491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,128,0.006365333166387346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,16384,0.021627555290857952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,3584,1.377527978685167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,2048,0.01407999959256914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,7168,0.018182223041852314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,256,0.039466665850745306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,3584,0.011104889214038849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,512,0.005631999837027655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,1024,0.004087999877002505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,768,0.0036880001425743103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,2048,0.007888889147175683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,768,0.18189688523610434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,256,0.005631999837027655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,2048,0.006221333311663733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,4096,0.010446222292052375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,2048,0.009873777627944946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,2560,0.005912000106440649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,7168,0.0199617776605818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,3584,0.01015911085738076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,768,0.03240888979699876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,3584,0.007747555772463481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,16384,0.008058666355080074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,2560,0.012168889244397482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,5120,0.028377778000301782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,8192,0.14110310872395834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,8192,0.20296888881259492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,64,0.003973333372010125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,3072,0.06827377610736422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,8192,0.15603555573357478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,3072,0.006633777585294511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,768,0.003724444243643019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,768,0.00499911109606425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,512,0.08455910947587754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,12288,0.017437333861986797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,12288,0.08476533492406209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,2048,0.013537777794731988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,32,0.005919999960396025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,64,0.002696000039577484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,7168,0.02270222206910451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,64,0.006711999989218182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,1536,0.010587555666764578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,256,0.013680000272062091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,5120,0.011236444115638733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,64,0.006224888894293044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,256,0.0064071110553211635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,16384,0.13436799579196507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,1024,0.039269334740108915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,6144,0.011459555890825061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,768,0.0074728892909155945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,3584,0.07458577553431193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,2048,0.00871377769443724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,5120,0.018193778064515855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,2560,0.006136000156402588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,3072,0.014177777700954013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,3584,0.017086222767829895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,10240,0.02886311213175456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,128,0.017616000440385606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,7168,0.015026667051845126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,5120,0.12359021769629584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,1024,0.006681777950790193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,64,0.0069439998931354946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,128,0.0034942221310403612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,2560,0.01864622202184465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,1024,0.004925333377387789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,2048,0.006672888994216919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,1536,0.009485333330101436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,16384,0.1536302169164022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,10240,0.03420622150103251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,2048,0.011751111182901593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,512,0.0074746666683091064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,3072,0.41257688734266496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,2560,0.03885866536034478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,1024,0.008220444122950235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,32,0.012125333150227865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,1024,0.0064035554726918536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,8192,0.03812711106406318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,32,0.008079999850855933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,768,0.7231653531392416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,8192,0.02197066611713833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,10240,0.014813333749771118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,3072,0.009242666264375051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,7168,0.02071555621094174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,4096,0.0063999998900625445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,12288,0.041309333509869046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,256,0.006437333507670297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,1024,0.005764444255166584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,16384,0.051627556482950844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,768,0.008059555457697975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,12288,0.08490755822923447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,512,0.006816000160243776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,5120,0.020770667327774894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,512,0.007301333049933116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,12288,0.019664888580640156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,2048,0.0058906666106647914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,4096,0.011605333122942181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,2048,0.013071111506885953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,7168,0.02350133326318529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,3072,0.013439111411571503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,12288,0.014288889037238227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,6144,0.016920000314712524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,3584,0.030425777037938435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,4096,0.0075662218862109715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,768,0.00591111100382275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,1024,0.04185333185725742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,65536,0.055431110991372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,768,0.007265778051482306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,12288,0.0173653331067827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,16384,0.09481510851118301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,64,0.004053333153327306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,8192,0.01423377792040507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,512,0.005482666608360078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,1024,0.0052622221410274506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,3584,0.007237333390447829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,512,0.005543111099137201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,6144,0.009541333549552495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,1536,0.06092888779110379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,64,0.002609777781698439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,1024,0.0041617775956789655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,64,0.0026906666656335196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,6144,0.010015111002657149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,256,0.009947555760542551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,16384,0.08220355378256904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,768,0.02795733345879449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,1536,0.017483555608325534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,65536,0.14682400226593018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,10240,0.017238222890430026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,3072,0.008175111479229396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,12288,0.09304000271691217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,2048,0.5549262364705404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,4096,0.009559111462699043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,4096,0.013415999710559845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,1024,0.003998222036494149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,7168,0.17623644404941133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,12288,0.007100444700982835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,2048,0.008832889298597971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,4096,0.014190221826235453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,768,0.007074666519959767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,7168,0.01127911110719045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,256,0.0034968890249729156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,1024,0.008568888737095727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,32,0.002557333351837264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,32,0.011445333560307821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,2560,0.1438408825132582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,5120,0.039956443839603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,512,0.009552000297440423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,3584,0.00797333319981893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,4096,0.010621333287821876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,16384,1.3792897330390081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,3584,0.009200000100665623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,2560,0.028723554478751287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,128,0.008448888858159384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,1536,0.005702222386995952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,65536,0.3857448895772298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,1536,0.0058906666106647914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,5120,0.011815999945004782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,2560,0.006239999913507038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,6144,0.005970666805903117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,8192,0.22066133552127412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,1024,0.003933333274390963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,2048,0.018176888426144917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,7168,0.27283912234836155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,512,0.003367111086845398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,32,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,64,0.0026871110830042097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,5120,0.013190222283204397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,3584,0.6990648905436198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,128,0.004965333475006951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,12288,0.019301333361201815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,10240,8.164517720540365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,7168,0.022853333089086745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,2560,0.019567999574873183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,2560,0.1214568879869249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,2048,0.03124622172779507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,4096,0.021112889051437378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,6144,0.1740800009833442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,2560,0.0074924445814556545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,512,0.0031191110610961914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,8192,0.02149244480662876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,12288,0.08031644423802693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,65536,0.46599732504950625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,128,0.008551999926567078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,8192,0.0747128857506646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,6144,0.009893333746327294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,32,0.05050666795836555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,1536,0.005960000058015187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,12288,0.018779555956522625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,256,0.013919111755159167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,65536,0.12415022320217556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,2048,0.007281777759393056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,6144,0.04193777839342753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,1024,0.003951999876234266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,128,0.002773333340883255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,8192,0.021556443638271753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,8192,0.03136355678240458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,3072,0.009536888864305284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,128,0.005170666509204441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,5120,0.008868444297048781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,3584,0.006844444407357111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,10240,0.01752088963985443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,512,0.0069217777086628815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,4096,0.05886222256554497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,1536,0.00433955548538102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,1024,0.009069333473841349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,12288,0.04814755585458544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,10240,0.01628533336851332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,256,0.00434044458799892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,3072,0.01647822227742937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,128,0.006299555715587404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,32,0.0064088888466358185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,65536,27.61968485514323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,256,0.015207999282413058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,6144,0.021896888812383015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,2560,0.014544000228246054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,5120,0.020788444413079154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,5120,0.009527111219035255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,512,0.005467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,4096,0.031769778993394636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,1536,0.05865066581302219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,2560,0.0063928887248039246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,3584,0.008719999757077958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,512,0.11461688412560357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,4096,0.03415022293726603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,1536,0.010535111029942831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,10240,0.17387821939256456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,4096,0.020288000504175823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,7168,0.04705511199103462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,8192,0.015417777829700045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,32,0.010451555252075195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,32,0.005761777775155173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,8192,0.05089422067006429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,32,0.01052977806992001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,1024,0.006609777609507243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,7168,0.01290933373901579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,64,0.003084444544381566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,128,0.0026906666656335196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,4096,0.010772444307804108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,512,0.003256888853179084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,7168,0.024879111184014216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,65536,0.5373555819193522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,65536,0.3396480083465576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,65536,0.04901688959863451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,3584,0.007677333222495184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,128,0.0032960000551409195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,65536,1.2395298216078017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,5120,0.08221866687138875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,6144,0.027913779020309448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,2560,0.019494222270117868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,65536,0.05568266577190823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,768,0.008762666748629676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,256,0.004301333179076512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,256,0.005439110928111606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,10240,0.13672622044881186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,10240,0.40311556392245823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,3072,0.026344888740115698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,32,0.004006222304370669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,512,0.005517333332035277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,5120,0.007135111424658034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,7168,0.014702222413486905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,65536,0.11284089088439941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,8192,0.040834665298461914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,2048,0.007304888632562425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,32,0.005842666659090254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,12288,0.02918133470747206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,768,0.005719111197524601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,12288,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,16384,0.06744266880883111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,2048,0.007094221810499827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,8192,0.019892444213231403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,6144,0.023556444380018447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,2560,0.011639110743999481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,6144,0.01828977796766493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,256,0.004111111164093018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,4096,3.543015162150065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,256,0.008911111288600499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,10240,0.13413066334194607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,7168,0.01610933409796821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,3072,0.009678222239017487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,1536,0.031535999642478094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,12288,0.016380444169044495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,10240,0.07608711057239108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,16384,0.12145955032772487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,4096,0.010379555324713389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,2048,1.8056115044487848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,65536,0.373016013039483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,1536,0.006777777853939269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,768,0.008362666600280339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,65536,0.07130310932795207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,768,0.005163555757866966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,4096,0.008368000388145447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,1536,0.006380444599522485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,2560,0.01385155568520228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,16384,0.3807831075456407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,7168,0.018938667244381376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,8192,0.007845333053006066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,1536,0.026774222652117412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,64,0.004934222333961063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,10240,0.013528889252079858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,10240,0.021196444829305012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,10240,0.006794666664467917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,6144,0.023751111494170293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,512,0.004956444518433677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,65536,0.08210400078031752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,768,0.00654666663871871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,16384,1.354456901550293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,1024,0.00508711114525795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,768,0.006663110935025745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,6144,0.017874666386180453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,1536,1.2557733323838975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,2560,0.007096889118353526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,128,0.005166222237878376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,128,0.015255999233987598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,12288,0.08012355698479547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,4096,0.007807999849319458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,6144,0.02091911103990343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,10240,0.038894222842322454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,65536,0.11066044701470269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,65536,0.19507111443413627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,256,0.00654755574133661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,3072,0.008787555827034844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,64,0.005141333159473208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,32,0.005373333477311664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,256,0.0029475556479560006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,10240,0.006581333362393909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,5120,0.024285333024130926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,4096,0.01804266704453362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,2048,0.013474666410022311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,16384,0.021539555655585393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,1024,0.009277332988050249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,6144,0.009903999666372934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,256,0.006162666612201267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,65536,0.06828177637524076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,1536,0.010856889188289642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,6144,0.06692088974846734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,65536,0.2284302181667752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,8192,1.7032018237643773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,32,0.0031822222388452957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,3072,0.08146488666534424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,3584,0.01779022150569492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,7168,0.012611555556456247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,3072,0.03626666797531976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,12288,0.03367822368939718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,6144,0.017608000172509086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,16384,0.08766133255428737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,512,0.006592888798978593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,768,0.0052177777720822235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,768,0.004004444513056013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,12288,0.02083200050724877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,4096,0.02475822303030226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,2048,0.006480000085300869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,16384,0.021504888931910198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,10240,4.253621419270833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,12288,0.04362933172119988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,2560,0.006575111299753189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,12288,0.018350222044520907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,3584,0.00628355559375551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,8192,0.014567999376191033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,3584,0.010989333192507425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,16384,0.06872089041603936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,256,0.006804444309737947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,4096,0.0251555558707979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,3072,0.01091733326514562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,32,0.006722666737106111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,6144,0.016343999240133498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,16384,0.037328001525666975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,1536,0.008466666771305932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,768,0.004648000001907349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,5120,0.009378666679064432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,2560,0.05364977651172214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,6144,0.05956266985999214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,8192,0.18039822578430176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,1024,0.008752000000741746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,2048,0.010081777969996134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,256,0.00535022219022115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,8192,0.03070666723781162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,7168,0.021439111895031396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,1536,0.013163555827405719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,7168,0.21910222371419272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,1024,0.004671999977694618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,512,0.010481778118345471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,2048,0.007567110988828871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,256,0.021800888909233943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,1024,0.007782222496138678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,1024,0.00498844434817632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,4096,0.006110222389300664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,5120,0.009865778187910715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,4096,0.036181333992216326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,128,0.011199110911952125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,10240,0.18074488639831543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,256,0.036474668317370944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,32,0.0026666666898462507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,2560,0.017507554756270517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,4096,0.011900444825490316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,128,0.002809777855873108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,512,0.007970666719807519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,16384,0.022223111655977037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,7168,0.027671999401516382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,4096,0.010969777901967367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,4096,0.022080888350804646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,3072,0.006790222393141852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,768,0.0038204445607132385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,3072,0.404189321729872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,1024,0.04705066813362969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,64,0.005473777651786804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,5120,0.02903199858135647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,16384,0.029858665333853826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,2048,0.0107022225856781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,1536,0.00870311094654931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,8192,0.012799111505349478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,256,0.008590222232871586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,128,0.003972444269392225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,2560,0.007126222054163615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,65536,0.906876458062066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,3072,0.0068488890926043195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,7168,0.006406222366624408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,32,0.007698666718271043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,4096,0.014724445011880664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,5120,0.026384888423813715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,3584,0.03561866614553664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,7168,0.13144800398084852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,256,0.004345777961942884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,16384,0.10361955563227336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,2560,0.09143822060690986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,10240,0.014939554863505893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,10240,0.02274311085542043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,2048,0.006463999963468975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,512,0.005417777846256892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,3072,0.6279670927259657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,6144,0.012412444584899478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,3072,0.010203555226325989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,512,0.016005333926942613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,2048,0.006302222195598815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,10240,0.052645332283443876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,5120,0.0125937776433097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,32,0.007056000332037608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,256,0.007162666983074612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,64,0.0029964444951878656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,128,0.004277333203289243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,7168,0.04671377936999003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,10240,0.3929902182685004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,1536,0.00702311098575592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,1536,0.006811555474996567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,128,0.006196444647179709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,5120,0.012233778006500669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,2048,0.006864000111818314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,768,0.005009777843952179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,768,0.007259555160999298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,5120,0.011799111134476133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,128,0.0058515556156635284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,3072,0.04253066579500834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,3072,0.04595822095870972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,1536,0.014836443795098199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,128,0.02119111186928219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,10240,0.02225333286656274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,16384,0.022026666336589392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,64,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,16384,0.051022221644719444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,768,0.0058755555914507965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,512,0.00600355573826366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,10240,0.013965333501497904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,7168,0.2809884548187256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,4096,0.0196106665664249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,3072,0.012483555409643384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,16384,0.1310702164967855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,64,0.003379555626047982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,6144,0.133206221792433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,64,0.0029057777590221832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,10240,0.035271111461851336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,1536,0.005709333138333426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,768,0.007207999626795451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,65536,0.04134044382307265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,3584,0.010608889162540436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,3584,0.011565333439244164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,65536,0.5821510950724283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,12288,0.06030933062235514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,4096,0.006450666735569636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,3584,0.011912888950771756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,1024,0.005799110978841782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,128,0.003960000144110786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,5120,0.4517519738939073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,10240,0.06199022134145101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,128,0.0042257776690853965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,512,0.005713777823580637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,3072,0.007546666595670912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,10240,0.048892445034450956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,2560,0.005682666682534748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,1536,0.006311111152172089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,7168,0.011047999891969891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,65536,0.15027733643849692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,2560,0.013530666629473368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,8192,0.01484444406297472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,65536,0.7551449139912924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,16384,0.019621334142155118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,128,0.004126222183307012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,6144,2.5099680158827042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,1024,0.007438222567240397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,5120,0.01996799972322252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,16384,0.02408711115519206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,6144,0.017874666386180453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,64,0.003023111157947116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,32,0.0032035555276605818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,2048,0.00590133335855272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,65536,0.024870221813519795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,8192,0.04545422063933479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,12288,0.040046221680111356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,7168,0.007807110746701558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,5120,0.006989333364698622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,768,0.004528000122971005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,6144,0.00814400033818351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,2560,0.013456000222100152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,7168,0.0229759994480345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,1024,0.004442666553788715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,1536,0.011990221838156382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,6144,0.43883466720581055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,8192,0.022611555125978258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,3072,0.009532444179058075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,512,0.005286222116814719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,64,0.048241777552498706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,12288,0.0720320012834337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,1536,0.004679999831649992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,768,0.0058000000814596815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,3584,0.008934222161769867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,12288,0.006392000036107169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,10240,0.023528888821601868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,8192,0.010241777532630498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,768,0.003592888928121991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,1536,0.006098666538794835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,2048,0.020015999674797058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,7168,0.010281778044170803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,10240,0.06537155310312907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,12288,0.01756622228357527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,768,0.006242666807439592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,5120,0.16921777195400664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,4096,0.029429333077536687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,65536,0.06967111428578694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,1024,0.007078222102589077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,65536,0.07315999931759305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,8192,0.08880977498160468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,3072,0.012421333127551608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,12288,0.13479199674394396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,64,0.0025928889711697898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,5120,0.00665244460105896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,2048,0.007616889145639207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,2560,0.015050666199790107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,5120,0.3674613369835748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,768,0.12536622418297663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,1024,0.015835555063353646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,16384,0.07329422235488892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,8192,0.006789333290523953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,12288,0.043191999197006226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,5120,0.00851733320289188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,4096,0.05209333366817898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,65536,0.0685724417368571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,32,0.006004444426960415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,5120,0.01917155583699544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,1536,0.005145777844720417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,64,0.005051555732885997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,1024,0.0038515557017591265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,7168,0.04170666800604927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,1024,0.04093866546948751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,3072,0.01977511081430647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,128,0.01370933320787218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,8192,0.02602133320437537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,4096,0.019343111250135634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,16384,0.1469306680891249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,4096,0.020824000239372253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,1536,0.00813244448767768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,64,0.002475555572244856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,8192,0.012072888513406118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,1024,0.013012444807423485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,7168,0.013728000223636627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,7168,0.028183112541834515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,10240,0.02939288814862569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,2048,0.004886222382386525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,768,0.005165333135260476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,6144,0.1649751133388943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,64,0.005340444544951121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,2560,0.008468444148699442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,256,0.0028728888266616394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,256,0.003145777723855443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,32,0.007822222179836696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,1024,0.22357779079013398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,10240,0.0219217770629459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,768,0.014874666929244995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,12288,0.02163644466135237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,3584,0.16675999429490831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,256,0.002903999967707528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,1536,0.056086222330729164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,3584,0.009904888768990835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,65536,0.04887911015086704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,3072,0.016932444439993966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,1024,0.01329955541425281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,3584,0.011386666860845355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,1536,0.006334222025341458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,3072,0.015212444795502557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,4096,0.023323555787404377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,5120,0.4470542271931966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,12288,0.03414577907986111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,1536,0.010498666928874122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,256,0.003113777687152227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,10240,0.02496799992190467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,16384,0.04982399940490723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,3584,0.008261333737108443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,256,0.005274666680230035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,4096,0.016640888320075143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,16384,0.06199644671546089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,2048,0.007303111255168915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,5120,0.020787555310461257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,6144,0.04232355621125963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,768,0.024846222665574815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,12288,0.006005333529578314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,2048,0.016195555528004963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,1536,0.0059004442559348206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,256,0.005711110929648082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,1024,0.004037333445416557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,12288,0.043078223864237465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,2048,0.006425777657164468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,8192,0.005984000033802456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,2560,0.007242666350470648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,8192,0.03327644533581204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,512,0.0057795556883017225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,7168,0.017955554856194388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,128,0.003039111072818438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,8192,0.020992888344658747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,3584,0.014290667242474027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,1024,0.0039813332259655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,10240,0.44595024320814347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,8192,0.011531555818186866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,12288,0.033012443118625216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,2560,0.017887999614079792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,6144,0.03801600138346354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,2560,0.005913777897755305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,1536,0.006623999940024481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,256,0.007277333074145847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,256,0.0053431110249625305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,10240,0.03529777791765001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,16384,0.059954669740464955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,4096,0.008061333662933773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,256,0.005186666631036335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,12288,0.0935315555996365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,512,0.0037084445357322693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,256,0.004460444466935264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,3584,0.023423110445340473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,3584,0.01204622205760744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,32,0.005468444277842839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,16384,0.03933688998222351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,7168,0.2191528876622518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,3584,0.014664888381958008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,1024,0.012163555456532372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,128,0.004394666602214177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,256,0.004091555459631814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,8192,0.0562302205297682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,5120,0.013539555172125498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,1536,0.036511109934912786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,128,0.003506666670242945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,6144,0.023760888311598036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,12288,0.04583377639452616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,2560,0.007879111501905654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,10240,0.02387911081314087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,10240,0.0064106666379504735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,7168,0.010855110983053843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,256,0.005596444424655702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,2560,0.2276995446946886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,128,0.0026515554636716843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,16384,0.03333510955174764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,512,0.004862222406599256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,1024,0.005411555369695027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,768,0.01367022262679206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,1536,0.0063928887248039246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,4096,0.012578667038016848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,16384,0.053173333406448364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,16384,0.09381600220998128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,10240,0.19329599539438883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,16384,1.666527960035536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,3072,0.012161778079138862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,512,0.0046471113132105935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,6144,0.10595999823676215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,1536,0.004955555415815777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,12288,0.11567821767595078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,512,0.005500444521506627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,12288,0.07294311126073201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,65536,0.0572684473461575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,1024,0.025801777839660645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,4096,0.011235555840863122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,7168,0.024912888805071514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,3072,0.02195466723706987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,1536,0.02015911042690277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,768,0.018108444081412423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,1536,0.005688888745175467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,2048,0.37864798969692653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,3072,0.020974222156736586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,3584,0.006707555717892117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,6144,0.007252444823582967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,768,0.004461333155632019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,256,0.021926222576035395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,3584,0.007245333658324347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,128,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,768,0.04223377837075127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,768,0.0035955554081334006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,256,0.006255110932721033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,256,0.005530666559934616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,1024,0.006179555422729916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,4096,0.010259555445777046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,2560,0.008011555506123437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,256,0.0031617778456873367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,128,0.005728000154097875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,1536,0.17036355866326225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,8192,0.032079110542933144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,16384,0.1817422178056505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,7168,0.05751199854744805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,2048,0.06124355395634969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,10240,0.022042666872342426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,768,0.008361777497662438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,32,0.002954666606254048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,128,0.006608888920810487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,128,0.0027111110587914786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,2560,0.007855999800893996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,8192,0.01943111088540819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,1536,0.004920888692140579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,64,0.0028088889602157804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,7168,0.01928444372283088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,1536,0.006396444307433234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,3072,0.01630311045381758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,3584,0.016072000066439312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,256,0.004971555537647671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,6144,0.012473777764373355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,768,0.012548444171746572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,12288,0.660204463534885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,6144,1.2374043994479709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,1024,0.0052266667286554975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,2048,0.04340711236000061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,1024,0.003620444486538569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,1536,0.004881777697139316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,4096,0.013850666582584381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,128,0.002982222164670626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,256,0.005116444494989184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,128,0.0027591110103660156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,64,0.08789066473642985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,768,0.008010666403505538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,512,0.0070168889231152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,768,0.04033689035309686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,6144,0.008239111138714684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,3584,0.008531555533409119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,256,0.021726222501860723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,6144,0.034318221939934626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,16384,0.19919821951124403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,1024,0.004850666556093427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,4096,0.015992000699043274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,5120,0.026148445076412622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,3072,0.018771555688646104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,65536,0.10955644316143459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,5120,0.009557333257463243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,16384,0.02966133422321743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,5120,0.02087733315096961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,2560,0.11199111408657497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,3072,0.0437466667758094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,16384,0.4526666535271539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,1024,0.004243555582231945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,8192,0.011331555744012197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,2048,0.005382222019963794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,2560,0.015455999308162265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,2048,0.006521777974234686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,65536,0.06612088945176867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,128,0.00526577772365676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,65536,0.06844711303710938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,10240,0.02071911096572876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,768,0.005236444373925527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,10240,0.01463733282354143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,128,0.0047235555118984645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,10240,0.028344889481862385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,4096,0.11142667134602864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,3072,0.10313955280515884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,128,0.045167999135123364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,32,0.021014221840434607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,8192,0.025234666135576036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,5120,0.012858666479587555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,7168,0.3201119899749756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,1536,0.007862221863534715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,65536,1.7268479665120442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,4096,0.007661333514584436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,16384,0.018814222680197824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,12288,0.03944177760018243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,256,0.0032844444115956626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,256,0.003020444470975134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,256,0.013944000005722046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,3584,0.012264000044928657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,768,0.008101333346631791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,64,0.030223998758527968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,2048,0.009651555783218807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,7168,0.011219555305110084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,10240,0.031423999203575984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,3072,0.09597688913345337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,16384,0.13369154930114746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,256,0.005234666582610872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,10240,0.09989955690171982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,3584,0.01069866700304879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,3072,0.007228444847795699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,5120,0.010000000397364298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,128,0.0029173334025674393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,1024,0.009303111169073317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,128,0.0069360000391801195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,512,0.003519111209445529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,1024,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,1024,0.0037004442678557504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,2560,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,768,0.006053333481152852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,512,0.006173333360089197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,3584,0.02646933330429925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,8192,0.017991110682487488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,2560,0.015198222464985318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,8192,0.011639110743999481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,10240,0.013295999831623502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,5120,0.012318222059143914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,768,0.003519111209445529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,512,0.01147822207874722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,768,0.005040888984998067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,3584,0.014486221803559197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,512,0.007483555210961237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,1024,0.009263111485375298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,1536,0.0063991112013657885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,64,0.01276000009642707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,12288,0.040208889378441706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,256,0.004869333157936732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,12288,0.02814399864938524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,7168,0.0064186664919058485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,6144,0.020038222273190815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,8192,0.0474586652384864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,1536,0.008953777452309927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,3584,0.008374222450786168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,128,0.005246222019195557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,6144,0.01088711122671763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,8192,0.018439110782411363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,32,0.005078222188684675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,3072,0.017118222183651395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,2560,0.05955466959211561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,12288,0.2643013265397814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,1536,0.006701333241330252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,64,0.007861333588759104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,3584,0.024104000793562994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,4096,0.012895111408498553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,2048,0.008784888519181145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,64,0.010128888818952773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,4096,0.008408000071843466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,768,0.005156444592608346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,6144,0.01294311136007309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,65536,0.25451911820305717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,128,0.0027919999427265594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,3584,0.012990222209029727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,1536,0.006320888797442119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,5120,0.008800889054934183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,32,0.002805333377586471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,16384,0.5010471343994141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,768,0.013207111093733044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,1536,0.004724444614516364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,7168,0.061872886286841504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,3072,0.007137777904669444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,1536,0.016007110476493835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,768,0.010696888797812991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,3584,0.0658044417699178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,3072,0.1520355542500814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,4096,0.01925155520439148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,12288,0.015167999598715039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,512,0.009016888837019602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,1536,0.00480355570713679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,512,0.01942311061753167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,1024,0.011539555258221097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,1536,0.006334222025341458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,6144,0.06620177957746717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,2560,0.0073520003093613525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,16384,0.03898577888806661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,16384,0.8089804649353027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,4096,0.010515555739402771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,1536,0.10001066658231948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,12288,0.0892711083094279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,2560,0.008476444416575963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,1536,0.009117333425415887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,5120,0.013723555538389417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,8192,0.013160000244776407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,4096,0.12788977887895372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,3072,0.03130933311250474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,3072,0.012689777546458773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,6144,0.2704239951239692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,2048,0.010096888575288985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,65536,0.2668444315592448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,32,0.005116444494989184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,6144,0.024777778320842322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,128,0.0052933332820733385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,1536,0.005635555419656966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,1024,0.011866666376590729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,8192,0.013148444394270578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,8192,0.29689155684577095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,1536,0.010031110710567897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,2048,0.051447110043631665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,16384,0.5639102194044325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,6144,0.0650800002945794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,256,0.00398577791121271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,256,0.0030542222989930045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,1536,0.005719999886221356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,7168,0.03497066762712266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,10240,0.06717422273423937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,65536,0.06571377648247613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,512,0.005409777578380372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,16384,0.0695448915163676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,3584,0.012694222231705984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,1536,0.005377777748637729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,1024,0.008522666990756989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,10240,0.05537155601713392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,65536,0.426508453157213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,16384,0.051933334933386914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,1024,0.006172444257471297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,12288,0.016467556357383728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,3584,0.010330666270520952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,12288,0.01907999979125129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,7168,0.017775999175177682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,256,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,7168,0.010794666906197866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,1536,0.011765333513418833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,3072,0.009029332962301042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,256,0.006607111129495833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,8192,0.2502044571770562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,16384,0.026985777748955622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,8192,0.010523555179437002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,4096,0.007920888562997183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,768,0.006574222197135289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,12288,0.01941866676012675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,10240,0.31093067593044704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,1024,0.00887111077706019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,512,0.006052444378534953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,1536,0.006673777682913675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,128,0.0063688887490166565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,5120,0.012263110942310758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,512,0.008650666309727563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,10240,0.17602311240302193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,3584,0.010659555594126383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,512,0.004927111168702443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,6144,0.014594667487674289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,32,0.0026488889836602737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,3072,0.01185955521133211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,65536,0.0291884442170461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,128,0.004407999830113517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,1024,0.005828444328573015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,1536,0.015988444288571674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,3584,0.01647644407219357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,32,0.0046453335218959385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,512,0.0041128889554076726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,5120,1.029280874464247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,7168,0.13705510563320583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,5120,0.02130488885773553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,256,0.0063795554969045855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,10240,0.020291555258962844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,6144,0.02574399941497379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,2048,0.006688888702127669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,4096,0.017439999514155917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,2560,0.006990222467316522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,2560,0.012423999607563019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,3584,0.16985510455237496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,2560,0.007098666495747036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,2560,0.00887288898229599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,3072,0.008896888958083259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,6144,0.01499466598033905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,8192,0.05340977840953403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,8192,0.01462933255566491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,256,0.011262222296661802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,128,0.005588444570700328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,3072,0.010976889067225985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,512,0.005387555393907759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,4096,0.009823111196359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,16384,0.10545955763922797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,32,0.006807999892367258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,2048,0.04231111208597819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,512,0.005768888940413793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,1024,0.06308888726764254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,7168,0.13160178396436903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,5120,0.03405244482888116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,12288,0.024000888069470722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,64,0.010116444693671333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,3072,0.007445333732499017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,65536,0.2183093362384372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,3584,0.017068444026841056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,512,0.0035333335399627686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,2048,0.007316444483068254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,5120,0.009278222090668148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,64,0.005220444252093633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,7168,0.15285600556267634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,16384,0.08825421995586819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,128,0.0027822222974565295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,256,0.026382222771644592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,3584,0.0374035570356581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,10240,0.013478221992651621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,7168,0.015186667442321777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,2048,0.0069982219073507525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,5120,0.006609777609507243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,2560,0.007174222005738153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,32,0.006773333168692059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,512,0.0047217777205838096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,2048,0.02772444486618042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,10240,0.046335998508665294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,16384,0.08258222209082709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,128,0.0027644443843099806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,768,0.005472888963090048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,65536,1.1565981970893013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,512,0.024663110574086506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,128,0.004111999852789773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,7168,0.04166400101449754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,5120,0.10020444128248428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,6144,0.03211022085613675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,8192,0.5041937828063965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,512,0.01074222226937612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,768,0.01331555595000585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,7168,0.01818933255142636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,16384,0.13248444928063288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,4096,0.007685333490371704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,5120,0.008826666408114964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,32,0.005187555733654234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,6144,0.014107555150985718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,768,0.0956488913959927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,256,0.005564444594913059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,1536,0.006467555546098285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,2560,0.022788445154825848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,3584,0.01757777730623881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,65536,0.4234791066911485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,7168,0.019935111204783123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,128,0.004243555582231945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,2560,0.17510310808817545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,768,0.008781332936551835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,2048,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,1024,0.004091555459631814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,32,0.004257777912749184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,12288,0.03923377725813124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,16384,0.027802666028340656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,4096,0.012245333029164208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,768,0.0058133333093590206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,65536,0.1005600028567844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,2048,0.007288888924651676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,7168,0.016180444094869826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,12288,0.2114266686969333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,65536,0.5540089077419704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,32,0.005386666705211003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,3072,0.06918399863772921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,2048,0.014307555225160388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,3072,0.011724444727102915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,4096,0.015051555302408008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,128,0.003248888792263137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,8192,0.019592000378502738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,512,0.008239111138714684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,3584,0.01705333259370592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,512,0.00666222224632899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,8192,0.006147555592987273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,2048,0.01668711172209846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,768,0.009950222240553962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,7168,0.01218577805492613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,512,0.005573333137565189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,10240,0.09067644675572713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,512,0.005008888741334279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,10240,0.06676089101367526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,2048,0.013728000223636627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,256,0.0064488889442549805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,2048,0.005387555393907759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,7168,0.06622222397062513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,4096,0.006059555543793573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,6144,0.011017777853541903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,7168,0.022082666556040447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,128,0.002642666714058982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,64,0.002612444468670421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,10240,0.013535111314720578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,3584,0.029672000143263075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,64,0.00739733295308219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,1024,0.0063279999627007385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,16384,0.02572088936964671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,128,0.006605333338181178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,2560,0.12470132774776882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,512,0.003656889001528422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,65536,2.254583994547526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,16384,0.045554667711257935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,6144,0.018372444642914664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,8192,0.015949333707491558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,16384,0.05273422267701891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,3584,0.13366844918992785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,8192,0.02857511242230733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,3584,0.01015022231472863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,512,0.018266666266653273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,16384,0.01832444469134013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,12288,0.021707556313938562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,256,0.005029333134492238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,7168,2.879872851901584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,64,0.003338666632771492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,2560,0.009175111022260454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,16384,0.41307467884487575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,1536,0.029224889145957098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,2048,0.1279768943786621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,2048,0.14374489254421657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,12288,0.007544889218277401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,5120,0.016936888297398884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,5120,0.01312711089849472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,7168,0.1229706605275472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,5120,0.07323733303281996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,768,0.004855999930037392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,65536,0.06609866354200575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,32,0.0053582220441765254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,2048,0.03265422251489427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,3584,0.015078221758206686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,512,0.0075031113293435835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,3072,0.028748446040683325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,8192,0.668876436021593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,1536,0.08741066853205363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,512,0.004635555462704765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,10240,0.014948444234000312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,1024,0.013270222478442721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,2048,0.00654666663871871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,12288,0.020280000236299302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,2560,0.009336888790130615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,768,0.005017777697907554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,3072,0.013156444662147097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,65536,0.1596648825539483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,3072,0.008117333054542542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,6144,0.016061334146393668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,10240,0.012974222501118978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,7168,0.07318133115768433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,16384,0.018629334039158292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,8192,0.01424088908566369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,7168,0.04404977626270718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,1536,0.007104888558387756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,10240,0.22285156779819063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,5120,0.07467022207048205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,5120,0.011393778026103973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,2560,0.008270222279760573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,10240,0.17833156055874297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,65536,9.804029676649305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,256,0.003960000144110786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,256,0.005459555735190709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,768,0.003576888806290097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,4096,0.01776622235774994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,12288,0.006400888992680444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,12288,0.011772444678677453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,1024,0.00906311141120063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,7168,0.013532444834709167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,1536,0.0073626670572492815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,10240,0.02276355524857839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,128,0.006248000181383557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,6144,0.01551644504070282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,2560,0.01365600029627482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,4096,0.013887999786270989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,128,0.005190222213665645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,2048,0.18329066700405547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,32,0.006933333145247565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,10240,0.03423822257253859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,16384,0.053824888335333936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,1024,0.005961777849329843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,1024,0.0047351109484831495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,768,0.006291555447710886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,1536,0.06352000104056464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,256,0.002805333377586471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,6144,0.014961777461899651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,3072,0.008512000242869059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,1536,0.0058559998869895935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,65536,0.0677057769563463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,128,0.0029608888758553397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,128,0.0032506665835777917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,32,0.0052764444715446895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,32,0.021993777818149988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,6144,0.00961511085430781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,256,0.021609778205553692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,16384,0.029359999630186293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,4096,0.010381333529949188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,3072,0.15747643841637507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,128,0.0030737777964936364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,10240,0.10630400313271417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,7168,0.06727288828955756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,8192,0.020243555307388306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,128,0.0028248888750871024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,6144,0.047637333472569786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,128,0.0028799999919202593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,65536,0.18270488580067953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,3584,0.011346666349305047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,64,0.008186666501892937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,6144,0.010242666635248397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,64,0.0052488889131281115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,64,0.007855999800893996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,4096,0.015504888362354703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,7168,0.010375111467308467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,256,0.00489688871635331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,2048,0.01569333341386583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,3072,0.009083555804358589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,3584,0.015433778365453085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,2560,0.009573332965373993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,7168,0.04634666774008009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,5120,0.037300444311565824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,2560,0.01776622235774994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,3584,0.015784889459609985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,2048,0.013139555851618448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,64,0.032319999403423734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,5120,0.03261511193381415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,8192,0.020491555333137512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,1536,0.02422844370206197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,512,0.005681777579916849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,12288,0.04761155446370443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,3072,0.0063484443558586975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,8192,0.029302222861184016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,256,0.005712000032265981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,16384,0.05670666694641113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,1536,0.05605689022276136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,16384,0.04145333170890808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,8192,0.006532444308201472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,256,0.005229333208666907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,8192,0.14702310827043322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,256,0.0034657776769664553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,3584,0.010965333216720156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,5120,0.028203556934992473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,32,0.0028088889602157804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,128,0.0027155555370781156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,2048,0.00562755556570159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,2048,0.013663999736309052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,512,0.008841777841250101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,128,0.002798222212327851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,10240,0.07215466764238146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,512,0.032615999380747475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,8192,0.04749422272046407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,16384,0.043177776866488986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,10240,0.2259368896484375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,2560,0.07407555315229628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,128,0.005107555538415909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,32,0.003015111097031169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,12288,0.02015911042690277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,512,0.08413955238130356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,8192,0.010811555716726514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,5120,0.02202755543920729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,4096,0.006401777681377199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,128,0.006053333481152852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,5120,0.09275466865963405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,8192,0.0810755557484097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,8192,0.02390577726893955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,3072,0.012124444047609964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,3584,0.5418151219685873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,8192,0.01239733315176434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,512,0.046422223250071205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,64,0.012836444709036084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,65536,0.14254932933383518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,5120,2.0587103101942277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,768,0.004903111192915174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,16384,0.06901866859859891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,65536,0.07667911052703857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,12288,0.12496356169382732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,1536,0.009630222287442949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,2560,0.00979200005531311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,10240,0.6913306448194715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,2048,0.006225777582989798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,2048,0.00702044450574451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,128,0.01719555589887831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,128,0.0040960001448790235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,2048,0.09940977891286214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,1536,0.0064497776329517365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,2560,0.006783999916579988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,2048,0.06471644507514106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,10240,0.040884445110956825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,4096,0.007795555724038019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,512,0.007617777420414819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,768,0.006221333311663733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,5120,0.01090755561987559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,65536,0.039297779401143394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,1024,0.003921777837806278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,1024,0.00582311095462905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,5120,0.02134311033619775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,64,0.00300177786913183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,768,0.00432977784011099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,768,0.006805333412355847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,3072,0.015455999308162265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,2048,0.021334222621387903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,65536,0.06556799676683214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,2048,0.007086222370465596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,10240,0.07947022385067411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,7168,0.02772088845570882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,10240,0.39048533969455296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,8192,0.011585777832402123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,5120,0.015255111787054272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,5120,0.00941244430012173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,12288,0.04579288760821024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,2560,0.006053333481152852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,32,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,32,0.011272889044549731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,16384,0.03193333413865831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,16384,0.06109866831037733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,7168,0.028159111738204956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,6144,0.057465778456793896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,32,0.00258577780591117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,16384,2.4966470930311417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,1536,0.012964444855848948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,2560,0.4938666555616591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,3072,0.009183111290136972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,5120,0.012903110848532783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,512,0.00850133349498113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,65536,1.7626373502943251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,256,0.006824888702895906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,2048,0.011183111204041375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,1024,0.010652444428867765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,128,0.004930666751331753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,1536,0.01312355531586541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,3584,0.017236444685194228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,4096,0.015112888481881885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,3584,0.6434497833251953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,3584,0.014945778581831189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,1024,0.09634221924675836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,1536,0.0058551111982928375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,32,0.00870577742656072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,32,0.00830488900343577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,2560,0.014947555131382413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,12288,0.5259813202752007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,16384,0.04381600022315979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,2048,0.00847999999920527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,2560,0.01590133375591702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,10240,0.056940442985958524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,5120,0.012168889244397482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,3072,0.05578844414816963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,16384,0.523131529490153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,65536,0.2648044427235921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,4096,0.048651556173960366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,128,0.018434666925006442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,512,0.008580444587601556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,512,0.004407999830113517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,2560,0.015270221564504834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,5120,0.014114666316244336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,4096,0.05194311009513008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,512,0.005054222212897407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,2560,0.042027556233935885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,2048,0.006643555644485686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,6144,0.010257778068383535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,2560,0.018556444181336295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,1536,0.00590133335855272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,2048,0.00979466653532452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,16384,0.026530666483773127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,8192,0.028844445943832397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,6144,0.01645155582163069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,12288,0.023678221636348303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,8192,0.2685146596696642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,65536,0.13381866614023843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,7168,0.1538355615403917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,2048,0.006597333484225803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,3584,0.030958222018347845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,7168,0.0982017781999376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,65536,3.2375965118408203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,16384,0.5228088696797689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,12288,0.022296888960732356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,16384,0.019877332780096266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,6144,0.05324622326427036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,512,0.010183999935785929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,16384,0.10057689083947076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,6144,0.01015733347998725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,1024,0.005508444375462002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,512,0.03775022096104092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,768,0.005344888816277186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,256,0.004544888933499654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,128,0.03007377849684821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,2048,0.1854542228910658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,3072,0.006918222126033571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,3072,0.01696622206105126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,3584,0.011751111182901593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,1024,0.011247110863526663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,2048,0.006792000184456508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,2048,0.05867999792098999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,3584,0.13930133978525797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,128,0.004193777839342753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,12288,0.02197955548763275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,2560,0.11559200286865234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,768,0.0035911111368073356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,5120,0.012062222593360476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,768,0.005124444348944558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,65536,5.685607062445746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,7168,0.029483556747436523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,32,0.08118755287594266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,2560,0.038471112648646034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,6144,0.054583999845716685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,6144,0.018208000395033095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,6144,0.07903377877341376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,6144,0.01759999990463257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,256,0.0029075555503368378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,7168,0.010590222146775989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,3072,0.15070933765835232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,768,0.008010666403505538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,1536,0.032447109619776406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,4096,0.04212444358401828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,12288,0.01569333341386583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,2560,0.009485333330101436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,10240,0.032391998502943255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,64,0.016295111841625638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,256,0.004558222161398994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,12288,0.28401954968770343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,32,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,6144,0.012536889149083031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,12288,0.01182755579551061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,10240,0.03460266523891025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,12288,0.06992444727155897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,1536,0.012273777690198688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,12288,0.09170222282409668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,2048,0.007035555938879649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,1024,0.005816000203291575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,2560,0.01291911138428582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,256,0.00720088928937912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,1536,0.006004444426960415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,2048,0.013257777525318993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,12288,0.03961599866549174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,64,0.006439111298984951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,4096,0.01978933314482371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,512,0.005591111050711737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,5120,0.01071733319097095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,32,0.0027626665929953256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,7168,0.05704533391528659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,512,0.0075493330756823225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,2048,0.015038222074508667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,16384,0.021448888712459143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,128,0.006114666660626729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,6144,0.03728444377581278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,4096,0.01862844493654039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,768,0.008475555313958062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,1024,0.005117333183685939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,10240,0.02966044346491496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,64,0.005110222018427319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,2560,0.007679111427730984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,1536,0.007249777515729268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,12288,0.06140977806515164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,4096,0.01753155555990007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,512,0.00535999983549118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,2048,0.007234666910436418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,6144,0.18819733460744223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,7168,0.02829688787460327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,7168,0.007090667055712805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,16384,0.10870755381054348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,768,0.01128266668981976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,768,0.005234666582610872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,128,0.027666666441493567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,65536,0.06953155332141452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,256,0.006543111056089401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,3072,0.09671377473407322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,1536,0.010203555226325989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,6144,0.02821777926550971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,768,0.007068444457319047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,3072,0.01943555474281311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,4096,0.02809600035349528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,8192,0.021133333444595337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,10240,0.024648000796635944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,512,0.009531555904282464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,1024,0.006063999815119638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,7168,0.011970666547616323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,3072,0.00776977754301495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,65536,0.31142844094170463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,5120,0.014979556202888489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,65536,0.05970755550596449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,5120,0.01861422260602315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,2048,0.010910222099887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,256,0.02665422194533878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,6144,0.012877333495352002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,2048,3.4255716535780163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,256,0.003689777933888965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,5120,0.011032000184059143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,65536,0.014479999740918478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,768,0.007122666471534305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,12288,0.11838400363922119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,7168,0.06430400080151029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,10240,0.31289421187506783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,128,0.006294222341643439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,2560,0.008108444511890411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,6144,0.014073777529928418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,64,0.002622222320901023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,1536,0.006375999914275275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,2560,0.012711111042234631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,512,0.004882666799757215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,3584,0.009713778065310584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,10240,0.04131822122467889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,8192,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,32,0.0026444445053736367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,2048,0.0053431110249625305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,4096,0.02476355599032508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,65536,0.36385244793362087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,3072,0.015764445066452026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,2048,0.03984444340070089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,512,0.007155555817815993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,10240,0.026187555657492742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,8192,0.006078222145636876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,512,0.00552977787123786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,5120,0.019179556104871962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,256,0.04540355669127571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,6144,0.2085040012995402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,32,0.42894310421413845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,32,0.003936888857020272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,32,0.00601244428091579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,768,0.02504088812404209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,3072,0.17295466528998482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,128,0.004120888809363048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,3072,0.008278222547637092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,768,0.020797333783573575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,16384,0.029764443635940552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,2048,0.12656177414788142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,64,0.003032888803217146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,256,0.00481599983241823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,1024,0.17217600345611572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,3072,0.025102222959200542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,1536,0.008253333469231924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,256,0.005223999834722943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,1024,0.11648444334665935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,6144,0.017416000366210938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,7168,0.0107351111041175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,1024,0.010244444840484196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,3584,0.3399226665496826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,2048,0.010911111202504901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,2048,0.009415999882751042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,32,0.003140444349911478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,65536,0.14954755041334364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,256,0.012378666963842181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,1536,0.011877333124478659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,2048,0.0053377776510185665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,64,0.002875555513633622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,65536,0.026438222991095647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,32,0.004664888812435998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,7168,0.016220443778567843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,10240,0.07380800114737616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,5120,0.024784889486100938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,128,0.00849333322710461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,8192,0.022876444790098403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,1536,0.011253332926167382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,7168,0.0064151109092765385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,2560,0.015205333630243937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,32,0.004826666580306159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,10240,0.3199555608961317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,512,0.009943111075295342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,2560,0.01512888901763492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,128,0.004229333251714706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,16384,1.041468408372667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,3072,0.009748444788985783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,6144,0.11691644456651475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,6144,0.013307555682129331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,2560,0.007674666742483775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,3072,0.011481777661376528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,3072,0.00793599999613232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,512,0.0033688888781600525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,12288,0.47537867228190106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,7168,0.01035822182893753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,1024,0.006851555572615729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,8192,0.052031109730402626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,2048,0.04872622092564901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,256,0.003064888839920362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,2048,0.038226667377683855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,512,0.006233777850866318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,1536,0.006315555423498154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,2560,0.013304000099500021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,6144,0.030210667186313208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,1024,0.0059573331640826324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,2048,0.007103111180994246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,512,0.0033653332955307434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,3584,0.008226667013433244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,16384,0.0960586667060852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,1024,0.007061333292060428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,6144,0.014088888963063559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,128,0.016209777858522203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,2560,0.011152000062995486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,2048,0.009415110780133141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,16384,0.023090667194790308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,16384,0.0850471125708686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,8192,0.33403200573391384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,3072,0.1118862231572469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,768,0.00544177782204416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,2560,0.009540444446934594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,16384,0.05573599868350559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,65536,0.2903262244330512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,7168,0.02473955518669552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,768,0.011111111276679568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,64,0.0027217778066794076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,2048,0.0074417781498697065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,128,0.005362666729423735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,7168,0.051015112135145396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,128,0.005616000129116907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,4096,0.09120711353090073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,1536,0.015827554795477126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,4096,0.04346399837070041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,2048,0.011605333122942181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,128,0.009500444763236577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,65536,13.814896477593315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,8192,0.01757244434621599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,256,0.024013333850436743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,128,0.010090666512648264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,512,0.006766222417354584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,2048,0.00609777785009808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,3072,0.02720622221628825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,1024,0.039289779133266874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,7168,0.028357333607143823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,4096,0.024694222542974684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,65536,0.12141866154140896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,3072,0.0069875551594628235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,32,0.014339556296666464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,6144,0.026177777184380427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,64,0.010025777750545079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,3584,0.029168887270821467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,1024,0.009561777942710454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,65536,0.2575004365709093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,7168,0.024479111035664875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,512,0.009441778063774109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,1024,0.006309333360857434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,256,0.06728444496790568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,5120,0.91412353515625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,2048,0.006303999986913469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,1536,0.007794666621420119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,32,0.0038773334688610504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,5120,0.01124533348613315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,10240,0.0422435568438636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,6144,0.015508444772826301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,128,0.02476088868247138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,4096,0.006280000011126201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,10240,0.02267644471592373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,3072,0.007932444413503012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,3584,0.009847111172146266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,8192,0.04006577862633599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,256,0.0033119999700122406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,7168,0.014692443940374585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,256,0.5321688652038574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,12288,5.300936804877387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,7168,0.02163911031352149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,128,0.006677333265542984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,256,0.03396000133620368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,16384,0.06506844361623128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,256,0.004833777745564778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,1536,0.05169333351982964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,6144,0.01425155500570933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,65536,0.04487199915779961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,7168,0.015479111009173922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,2048,0.07868622408972846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,1536,0.00924800005224016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,12288,0.051701333787706166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,3072,0.006616888774765863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,64,0.003568888952334722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,6144,0.014010666145218743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,3584,0.010001777774757808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,3584,0.03477599885728624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,8192,0.029657777812745836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,4096,0.01165600038237042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,128,0.004791999856630961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,5120,0.022240888741281297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,65536,0.21243733829922148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,10240,0.08542222446865505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,10240,0.014746665954589844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,5120,0.009400000174840292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,7168,6.033030192057292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,6144,0.04457688993877835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,32,0.002887111157178879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,256,0.0036142224238978494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,16384,0.034605334202448525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,3584,0.013603555659453074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,1536,0.014377777775128683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,2560,0.00591822216908137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,65536,0.07093777921464708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,512,0.004200889004601372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,6144,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,7168,0.0905288921462165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,2560,0.008182221816645728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,64,0.0027715555495686005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,2048,0.007153777612580194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,5120,0.06623644298977323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,5120,0.01644533375898997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,1024,0.00925866680012809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,1024,0.009917333722114563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,1024,0.007536888950400882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,1024,0.12330667177836101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,4096,0.008647999829716152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,12288,0.024471999870406255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,1024,0.00434311106801033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,3584,0.02699111070897844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,7168,0.01923644377125634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,2560,0.007848000360859765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,5120,0.17080356015099418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,2560,0.006711111300521427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,32,0.0025964443468385274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,8192,0.02089599933889177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,1024,0.005714666512277391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,128,0.0028826666788922418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,10240,0.18636355135175917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,4096,0.008756444685988957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,8192,0.043748444981045194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,6144,0.006616888774765863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,10240,0.010743111371994019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,768,0.004813333352406819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,16384,18.71722920735677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,128,0.005404444618357553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,6144,0.06395822101169162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,512,0.007623111208279927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,512,0.006517333288987477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,4096,0.012587555580668978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,3072,0.006582222051090664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,65536,0.05380444394217598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,6144,1.4310951232910156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,7168,0.019426667028003268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,3072,0.010830221904648675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,10240,0.04350489046838549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,128,0.009376000199053023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,3584,0.44412000974019367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,256,0.032109333409203425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,3584,0.01629155543115404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,768,0.08360177940792507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,1536,0.005048888838953442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,65536,0.033744888173209295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,2048,0.007239110767841339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,65536,0.2818782329559326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,65536,0.015174221661355762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,128,0.026005332668622334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,3072,0.022807111342748005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,7168,0.019675556156370375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,1536,0.00758133331934611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,32,0.00518488883972168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,8192,0.03066666589842902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,32,0.005208000126812193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,4096,0.07098311185836792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,12288,0.006313777632183499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,65536,0.22383112377590605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,4096,0.010472888747851053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,5120,0.0218542218208313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,8192,0.011711111499203576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,3584,0.019738665885395475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,128,0.003915555361244413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,6144,0.048801776435640126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,7168,0.013363555901580386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,128,0.004398222184843487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,16384,0.021006221572558086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,8192,0.04313333498107063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,256,0.002842666581273079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,256,0.01962488889694214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,3072,0.007793777518802219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,1024,0.006120000034570694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,128,0.003212444484233856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,2048,0.028265777561399672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,7168,0.03481066558096144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,768,0.005184000151024925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,3584,0.008789333204428354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,2560,0.007347555624114142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,128,0.004040889028045866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,8192,0.01352355546421475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,5120,0.013194666968451606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,4096,0.009460444251696268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,4096,0.033949332104788885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,128,0.004301333179076512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,1536,0.0063279999627007385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,8192,0.02312711046801673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,128,0.014200000299347771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,4096,0.3171119954850939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,10240,0.17508533265855578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,256,0.008416888614495596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,32,0.004519111166397731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,32,0.00452888881166776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,128,0.00618577789929178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,5120,0.008492444124486711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,65536,0.17955733670128715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,256,0.005676444619894028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,768,0.021335111724005804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,7168,0.04570133487383524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,7168,0.006259555617968242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,6144,0.020650666620996263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,128,0.004156444635656145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,65536,0.10997066895167033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,65536,0.21706133418613008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,3072,0.0562266641192966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,2048,0.005220444252093633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,6144,0.013414222333166333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,1024,0.008223111430803934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,16384,0.28847911622789174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,6144,0.031194667021433514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,7168,0.010800888968838586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,1024,0.00572533326016532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,4096,0.009764444496896531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,7168,0.005906666732496685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,3072,0.007518221934636434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,7168,0.05502133236991035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,4096,0.01825066738658481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,4096,0.010444444086816577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,2048,0.006413333531883027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,1024,0.004941333499219683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,1536,0.008533333738644918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,6144,0.024234667420387268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,2560,0.011424889167149862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,3072,0.02462133268515269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,65536,0.1436577770445082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,64,0.007149332927332983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,256,0.022639999787012737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,32,0.016892444756295946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,5120,0.010242666635248397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,5120,0.24229688114590117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,8192,0.01822933389080895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,64,0.0052871112194326185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,1024,0.0076657773719893555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,768,0.005674666828579373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,6144,0.3769422107272678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,12288,0.02012355625629425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,768,0.007575111256705389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,1024,0.0064151109092765385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,10240,0.01352355546421475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,64,0.01370488852262497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,10240,0.027520888381534155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,3584,0.009054222040706212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,512,0.00639111093348927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,256,0.02463111115826501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,6144,0.5448186662462022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,8192,0.016809778081046212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,2048,0.8616551293267144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,512,0.0876622200012207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,4096,0.08816799852583145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,10240,0.007986666427718269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,2048,0.007409777906205919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,6144,0.04195288817087809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,128,0.004274666723277834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,256,0.05695733096864489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,65536,0.09200711382759942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,12288,0.01426666643884447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,3072,0.020784889658292133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,10240,0.02683555583159129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,2560,0.007977777885066139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,3584,0.20327377319335938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,512,0.009123555488056606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,16384,0.022663111488024395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,768,0.016992888516849942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,512,0.014249778456158109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,4096,0.03588533401489258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,6144,0.030478222502602473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,4096,0.008584889272848764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,1024,0.008107555409272512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,1536,0.02419288953145345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,7168,0.053928001059426196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,12288,1.2982471254136827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,1024,0.007357333269384172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,2048,0.1090080009566413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,128,0.005670222143332164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,4096,0.08070400026109484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,64,0.008772444393899705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,32,0.0053839998112784485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,12288,0.009011555876996782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,256,0.0028782222006056043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,64,0.0027866665687825945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,65536,0.06512710783216688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,7168,0.022060443957646687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,12288,0.03602933221393161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,10240,0.02442933287885454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,16384,0.021737777524524268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,7168,0.01347555551264021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,4096,0.018983999888102215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,2560,0.0736266639497545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,16384,0.023776888847351074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,2048,0.01369866645998425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,32,0.0034142221427626083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,5120,0.006158222340875202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,12288,0.031311998764673867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,128,0.0029191111938820947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,16384,0.009740444521109263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,2048,0.04512622290187412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,65536,0.08048178089989556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,2048,0.015593777100245157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,768,0.010954666468832227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,3584,0.007743111087216272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,7168,0.029602666695912678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,16384,0.00656444455186526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,32,0.005048000150256687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,128,0.002833777831660377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,768,0.00535999983549118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,16384,0.049636443456014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,3584,0.03584622343381246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,3072,0.009887999958462184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,1536,0.08354310856925116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,10240,0.01885688967174954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,16384,0.03646399908595615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,1536,0.047817776600519814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,7168,0.2429431014590793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,10240,0.014392889208263822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,1024,0.01037333326207267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,768,0.013236444857385425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,128,0.037251555257373385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,7168,0.007278222176763747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,65536,0.08076000213623047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,3584,0.021529777182473078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,2560,0.005934222290913264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,65536,0.009333333207501305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,10240,0.03467199868626065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,3584,0.019512888458040025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,10240,0.01811733345190684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,256,0.004477333277463913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,16384,0.05793688694636027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,512,0.0033004445334275565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,65536,0.07221688826878865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,3584,0.030398223135206435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,2560,0.007119999991522894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,10240,0.013576889203654395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,12288,0.10537422365612453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,768,0.007748444047239091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,4096,0.0463084446059333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,1536,0.009760888914267222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,3584,0.007677333222495184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,10240,0.03388799892531501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,1536,0.01094044413831499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,512,0.004552888787455029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,6144,0.018208889497650992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,1024,0.004101333518822988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,768,0.0041999999019834726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,10240,0.029771556456883747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,5120,4.532617780897352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,6144,0.02850488821665446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,128,0.003989333493842019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,8192,0.01997066703107622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,128,0.004111999852789773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,1536,0.023166222704781428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,8192,0.01684266659948561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,10240,0.018242667118708294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,6144,0.08519911103778416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,1024,0.0053119998839166425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,2560,0.08803377548853557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,5120,0.017048888736300997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,1536,0.006790222393141852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,8192,0.0090844440791342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,10240,0.47915999094645184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,5120,0.012141333685980903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,2048,0.010681777364677854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,3072,0.12057777245839436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,8192,0.011713777979214987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,64,0.003085333233078321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,512,0.005487111293607288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,1024,0.010296888649463654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,32,0.006627555522653792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,16384,0.25466399722629124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,512,0.0033760000434186724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,2048,0.006353777729802662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,7168,0.013975999421543546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,32,0.0028862222615215513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,16384,0.010804444551467896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,8192,0.25176000595092773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,1024,0.008456889126035903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,2560,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,3584,0.008203555312421586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,768,0.09453244341744317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,6144,0.010144888526863521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,6144,0.05526399943563673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,512,0.007086222370465596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,3072,0.007241778075695038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,3072,0.010709332923094431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,768,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,1536,0.006478222293986215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,512,0.0041475556790828705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,512,0.0038933331767717996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,32,0.002473777780930201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,2048,0.10638222429487441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,2560,0.007368889119890001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,512,0.005377777748637729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,10240,0.32731466823154026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,1024,0.005160000175237656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,2560,0.007316444483068254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,16384,0.020949333906173706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,5120,0.04955910974078708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,1024,0.011751111182901593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,12288,0.06364355484644572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,256,0.004230222354332606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,65536,0.29251376787821454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,768,0.02370044423474206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,5120,0.011628444823953839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,3072,0.2413724528418647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,7168,1.445659531487359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,4096,0.022983110613293115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,256,0.006631999793979857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,4096,0.007287110719415877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,32,0.006071110980378256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,768,0.006522666662931442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,6144,0.02643377747800615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,3584,0.07193066676457723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,768,0.008640000389681922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,10240,0.03042666779624091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,256,0.009534222384293875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,512,0.004966222163703707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,6144,0.00983911090426975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,768,0.008101333346631791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,768,0.009411555197503831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,128,0.008639111287064022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,512,0.005039111193683413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,3072,0.016523554921150208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,2048,0.018353778454992507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,1024,0.005436444448100195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,3072,0.015196444259749519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,512,0.012815111213260226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,8192,0.028753777345021565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,32,0.006590222318967183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,512,0.011603555745548673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,32,0.0027679999669392905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,512,0.006263111200597551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,128,0.0028604444944196274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,2048,0.016186666157510545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,3072,0.013597333596812354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,128,0.004238222208287981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,8192,0.006711999989218182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,8192,0.01648888819747501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,128,0.004889777965015835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,65536,0.3701813220977783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,8192,0.1979831059773763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,7168,0.13754844665527344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,10240,0.013577777478430005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,3072,0.03271111183696323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,64,0.0025840000145965153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,3072,0.006322666588756774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,512,0.004657777647177379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,32,0.005872888697518243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,65536,0.06859733661015828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,2048,0.04265066650178698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,1024,0.005111999809741974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,32,0.003104000041882197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,10240,0.35971911748250324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,3584,0.01844888925552368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,4096,0.011767110890812345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,16384,0.023090667194790308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,2560,0.007177777588367462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,256,0.0074133334888352295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,4096,0.007801777786678738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,3584,0.017747556169827778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,16384,0.01849688920709822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,1536,0.00720355576939053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,16384,0.02316711180739933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,5120,0.021677333447668288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,16384,0.026147555973794725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,5120,0.16317866908179388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,1024,0.005460444423887465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,65536,0.06705600023269653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,64,0.005379555539952383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,65536,0.28725867801242405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,2048,0.005362666729423735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,3072,0.015259555644459195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,12288,0.009866666462686326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,5120,0.014684443672498068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,1024,0.007831999825106727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,512,0.005597333527273602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,1536,0.040771553913752236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,1536,0.011231111155615913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,5120,0.07105777661005656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,4096,0.03029866682158576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,1024,0.021685333715544805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,256,0.00423200014564726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,16384,0.14841066466437444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,6144,0.09973422023985122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,7168,0.013349333571063148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,12288,0.03050311075316535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,16384,0.009831111464235518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,3072,0.007174222005738153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,16384,0.03973244296179877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,5120,0.09056533045238918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,6144,0.010040000081062317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,6144,0.009910222556855943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,12288,0.054730667008293994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,65536,0.13507289356655544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,32,0.005123555660247803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,7168,0.010989333192507425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,32,0.004567999806669024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,64,0.0052720002002186244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,10240,0.08688000175688003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,5120,0.01456622282663981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,2560,0.012454222473833295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,256,0.00442844463719262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,1536,0.004658666749795278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,1536,0.007063999772071838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,7168,0.006867555694447623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,32,0.005151111218664381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,2048,0.019518221418062847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,5120,0.01692800058258904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,12288,0.26608088281419545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,7168,0.06950310866038005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,3072,0.02365600069363912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,5120,0.01575911045074463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,16384,0.05259911219278971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,7168,0.01698666645420922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,1536,0.3135928842756483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,64,0.0025884444928831523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,128,0.019840000404251944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,2560,0.07034666670693292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,768,0.012967999610635968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,8192,0.2509244547949897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,6144,0.025489777326583862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,3584,0.011801777614487542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,65536,0.5338960223727757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,2560,0.007344889144102733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,12288,0.024951110283533733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,2560,0.006608888920810487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,16384,0.08103733592563205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,64,0.0027111110587914786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,10240,0.048503110806147255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,768,0.005482666608360078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,16384,0.060770663950178355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,12288,0.036405334870020546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,1024,0.006807111203670502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,3584,0.01202311118443807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,16384,0.13475999567243788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,3584,0.008090666598743862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,3584,0.013739555246300168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,16384,0.018394667241308425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,768,0.006205333189831839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,1536,0.05658044417699178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,256,0.003847111016511917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,5120,0.041215999258889094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,64,0.0028968888024489083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,6144,0.010472888747851053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,1536,0.007374222079912822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,65536,0.17795199818081328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,5120,0.010199999643696679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,2048,0.032864888509114586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,1024,0.005487111293607288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,32,0.005452444569932089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,32,0.007735110819339752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,32,0.005255110975768831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,1024,0.004030222280157937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,7168,0.31058933999803334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,6144,0.008268444074524773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,128,0.0029022221763928733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,1024,0.006065777606434292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,1536,0.009124444590674506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,16384,0.09533244371414185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,32,0.04653955499331156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,3072,0.010422222316265106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,512,0.06316622098286946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,64,0.0028364445186323593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,7168,0.018034666776657104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,7168,0.03651555710368686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,3072,0.017097777790493436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,3584,0.023670223024156358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,16384,0.09619288974338108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,128,0.006598222172922558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,128,0.003916444463862313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,65536,0.5657084253099229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,6144,0.01904177831278907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,1536,0.011941333611806234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,10240,0.0331928895579444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,1024,0.008449777960777283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,512,0.003419555516706573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,3072,0.011297778122954898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,1536,0.009447111023796929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,3072,0.007941332956155142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,12288,0.25303732024298775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,1536,0.0221395558781094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,7168,6.492986467149522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,1536,0.010649777948856354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,2048,0.010316444767846001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,65536,0.09488266706466675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,1536,0.015538665983412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,5120,0.025842666625976562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,768,0.003654222107595868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,256,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,512,0.013201778133710226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,7168,0.22188711166381836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,512,0.026360889275868733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,256,0.00489777781897121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,8192,0.14794311258527967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,128,0.004106666478845808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,1536,0.14911644988589817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,5120,0.023707555400000677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,7168,0.010374222364690568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,8192,1.156926261054145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,32,0.007599111232492659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,16384,0.16569066047668457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,768,0.005508444375462002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,12288,0.015788444214397006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,1024,0.005678222411208683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,64,0.0025111111915773815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,10240,0.019702222612169053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,32,0.02735288937886556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,128,0.006215999937719769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,7168,0.01314311143424776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,3584,0.00904177791542477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,256,0.009383999639087254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,7168,0.011277333729796939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,512,0.005953777581453323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,16384,0.051345778836144336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,10240,1.1684000227186415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,1536,0.0074720001882976955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,65536,0.19591111607021758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,16384,0.006657777975002925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,65536,0.08119022183948092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,65536,0.6106693479749892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,2048,0.031119111511442397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,16384,0.046613332298066884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,2048,0.028446220689349707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,512,0.004209777961174647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,512,0.027512888113657635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,12288,0.3903484344482422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,8192,0.007631999750932057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,2560,0.009478222164842818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,128,0.010538666612572141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,8192,0.018406222263971966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,256,0.0048177776237328844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,65536,0.07471555471420288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,768,0.005210666606823604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,4096,0.009350222018029954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,64,0.0026853332916895547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,2560,0.011369778050316704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,12288,0.014091556270917257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,3584,0.00886933339966668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,10240,0.013658666776286231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,256,0.01774577796459198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,6144,0.09229777918921576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,2048,0.024409777588314478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,5120,0.006723555425802867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,256,0.00508088908261723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,2560,0.00868711123863856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,768,0.0035786665976047516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,4096,0.011234666738245221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,2560,0.011811555259757571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,768,0.051593777206208974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,16384,0.03539289037386576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,8192,0.04708355665206909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,10240,0.5688391261630589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,64,0.00847999999920527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,64,0.0064088888466358185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,16384,0.006610666712125142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,65536,0.10454222228791978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,3072,0.009937778115272522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,512,0.02132533325089349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,8192,0.011743110915025076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,768,0.007721777591440413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,10240,0.023817777633666992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,32,0.03863644599914551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,768,0.009306666751702627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,6144,0.0064177778032090925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,3072,0.22209600607554117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,1536,0.056384887960222035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,5120,0.01625155574745602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,256,0.003752888904677497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,7168,0.023404444257418316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,7168,0.05903200308481852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,6144,0.037251555257373385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,16384,0.023842665884229872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,16384,4.707054985894097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,5120,0.019919999771647982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,6144,0.010323555933104621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,65536,0.23808712429470488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,2560,0.009355555805895064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,5120,0.03629066546758016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,2560,0.00889511075284746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,768,0.06792889038721721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,3072,0.00872444444232517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,6144,0.012744888663291931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,12288,0.02997066577275594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,16384,0.049489776293436684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,12288,0.050168000989490084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,1024,0.003638222399685118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,64,0.005054222212897407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,6144,0.011219555305110084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,3584,0.013464888764752282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,16384,0.026734222968419392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,7168,0.06431911389033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,6144,0.08364444308810765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,1024,0.0051857779423395796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,3584,0.011721777419249216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,1024,0.006033777776691649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,256,0.010778666370444827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,3072,0.01556622154182858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,6144,0.03467999895413717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,5120,0.0657048887676663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,8192,0.04489955637190077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,512,0.005667555663320754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,6144,0.014190221826235453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,3072,0.007714666426181793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,1536,0.008808888494968414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,512,0.007041778001520369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,10240,0.19707911544375947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,65536,0.03204088740878635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,16384,0.3097075621287028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,1536,0.007496000164084964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,12288,0.019521777828534443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,12288,0.024031110935741003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,32,0.00692622239391009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,1024,0.049991110960642494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,3072,0.02459555533197191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,3584,0.015986666083335876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,3072,0.035022222333484225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,7168,0.07853244410620795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,512,0.0032417778339650896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,768,0.06487822532653809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,10240,0.19875910547044542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,10240,0.027936001618703205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,256,0.005091555416584015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,12288,1.1827306747436523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,12288,0.04996622271007962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,10240,0.03786755601565043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,7168,0.012341332932313284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,768,0.010662222074137794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,5120,0.01796888808409373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,64,0.006182222316662471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,3584,0.16189155313703749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,2048,0.005438222239414851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,12288,0.03391111228201125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,128,0.007900444169839224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,5120,0.026208889153268602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,3584,0.007292444507280986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,1536,0.015332443846596612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,8192,0.0945697757932875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,12288,0.21958843866984049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,3584,0.018375110295083787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,2560,0.014197332991494073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,16384,0.019312888383865356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,64,0.010189333723651038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,2048,0.005383111122581694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,3584,0.13951288329230413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,2048,0.0064275554484791225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,5120,0.017864000466134813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,1024,0.025437333517604407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,32,0.005819555371999741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,4096,0.011719110939237805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,32,0.00526577772365676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,8192,0.01423111061255137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,128,0.003875555677546395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,32,0.030792888667848375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,3584,0.021576000584496394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,10240,0.018094221750895183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,65536,0.21099289258321127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,8192,0.015189333094490899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,1536,0.0058559998869895935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,3584,0.3605875439114041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,128,0.003843555433882607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,256,0.003814222084151374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,16384,3.8746045430501304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,3584,0.019634667370054457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,5120,0.01683999929163191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,4096,0.1514915492799547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,12288,0.00721777809990777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,3072,0.023740443918440077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,2560,0.039572444227006696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,1024,0.006433777925040986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,2048,0.006870222174459034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,128,0.0029271110478374693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,768,0.02256800068749322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,128,0.03953511185116238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,4096,0.07137333022223578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,10240,0.0468408895863427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,3584,0.05018577641910977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,5120,0.009135999613338048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,4096,0.008043555749787224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,65536,0.030510223574108545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,4096,0.017690666847758822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,16384,0.03013955553372701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,128,0.002830222249031067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,768,0.007992888490358988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,3584,0.006610666712125142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,256,0.0028897778441508612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,256,0.004559111264016894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,12288,0.022999111149046157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,64,0.015875554747051664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,512,0.0033822223130199644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,1536,0.010583110981517367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,768,0.005175999883148405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,32,0.008096888661384583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,12288,0.019144889381196763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,64,0.045718222856521606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,12288,0.04670400089687771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,8192,0.0064702220261096954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,6144,0.02321333355373806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,768,0.05190044310357836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,16384,0.07168266508314344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,128,0.004436444491147995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,6144,0.05340533455212911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,3072,0.021720000439220007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,7168,0.011133333047231039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,512,0.005483555710977978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,768,0.005928888916969299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,64,0.002623999905255106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,3584,0.01018311083316803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,3072,0.05723110834757487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,6144,0.01147466649611791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,1536,0.019533332851197984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,65536,0.013636444177892474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,1024,0.00906044410334693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,1536,0.008526222573386298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,64,0.0032026666320032547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,10240,0.00976533359951443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,1024,0.1270293394724528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,1024,0.08811022175682916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,3072,0.006683555742104848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,256,0.40560089217291934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,12288,2.5307990180121527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,3072,0.010293333066834344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,2048,0.014338667194048563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,6144,0.029711110724343195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,10240,0.03139288889037238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,5120,0.011225777367750803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,16384,0.06204977962705824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,3584,0.04390933447413974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,4096,0.01813155578242408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,768,0.009956444303194681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,2560,0.014673777752452426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,768,0.007681777907742395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,3072,0.012262221839692859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,16384,0.2051724460389879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,3584,0.06774044699139066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,128,0.0030177777840031516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,16384,0.1157191064622667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,128,0.01646133263905843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,2560,0.016007999579111736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,3072,0.019480889042218525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,128,0.005003555367390315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,65536,0.08027288648817274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,3072,0.006828444285525217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,64,0.022435555855433147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,512,0.013169777890046438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,32,0.04253333475854662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,10240,0.05221066541141934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,16384,0.03423733181423611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,2560,0.006031999985376994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,512,0.004499555461936527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,10240,0.02869066596031189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,65536,0.008198222352398766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,12288,0.016791999340057373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,32,0.005136888888147143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,6144,0.009939555492666032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,3072,0.025013332565625507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,768,0.008582221964995066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,16384,0.5882773399353027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,7168,0.2193333307902018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,3072,0.04591822293069628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,8192,0.04552533229192098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,3584,0.05093510945638021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,12288,0.07691199911965264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,6144,0.009557333257463243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,512,0.0032924444725116095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,16384,0.22482222980923125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,64,0.006309333360857434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,128,0.005590222362014983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,512,0.003572444534964032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,12288,0.4535439809163411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,65536,1.8320816887749567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,8192,0.35811911688910586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,2048,0.006568000134494569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,6144,0.012116444607575735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,10240,0.03274310959710015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,32,0.002551111082235972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,5120,0.011569778124491373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,7168,0.012431110772821637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,2560,0.023568888505299885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,512,0.0064560001095136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,16384,0.30730756123860675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,768,0.01310755560795466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,10240,0.3068506717681885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,768,0.003474666840500302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,128,0.006172444257471297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,7168,0.015178667174445258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,64,0.0026453334010309647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,768,0.00646666685740153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,16384,0.015102222561836243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,2560,0.006515555497672822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,1024,0.0058764442801475525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,65536,0.09517866373062134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,6144,0.020424000091022916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,3584,0.006244444598754247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,1536,0.017607111069891188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,3072,0.3021750979953342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,10240,0.030062221818500098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,128,0.0038746665749284956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,8192,0.03730222251680162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,64,0.01763911048571269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,768,0.006134222365087933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,16384,0.026288888520664636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,3584,0.008486222061845992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,2048,0.010513777534166971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,5120,0.008928000099129146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,1536,0.005382222019963794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,12288,0.01015644437736935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,2048,0.0120319997270902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,64,0.0053137776752312975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,2560,0.01366666704416275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,64,0.002410666603181097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,32,0.005024888863166173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,2560,0.08501688639322917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,1536,0.005208000126812193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,1536,0.004804444395833545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,65536,0.22975556055704752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,10240,0.03712622324625651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,6144,0.02442933287885454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,2048,0.0064195555945237475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,4096,0.019664888580640156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,5120,0.028582221931881372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,5120,0.05013244350751241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,7168,0.03092799915207757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,1536,0.012653333445390066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,3584,0.08059466547436185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,4096,0.020555555820465088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,1536,0.011344888971911537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,768,0.018087999688254464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,2560,0.022806222240130108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,64,0.006117333140638139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,65536,0.022514666120211285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,16384,0.18191022343105742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,2560,0.016041777200169034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,12288,0.0999422205819024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,256,0.01534400052494473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,12288,0.015171556009186639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,16384,0.04639910989337497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,64,0.022583999567561682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,12288,0.03770133190684848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,1024,0.01698933376206292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,12288,0.13503911760118273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,7168,0.48833868238661027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,12288,0.023198222120602924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,768,0.004775111046102312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,1024,0.011687999798191918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,1536,0.005416000054942236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,3072,0.007152888509962294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,6144,0.03453244434462653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,3072,0.006093333164850871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,256,0.004873777843183941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,1024,0.006048888795905643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,768,0.028745777077145044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,8192,0.0626408921347724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,8192,0.018552000323931377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,7168,0.025184889634450276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,768,0.06425066788991292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,12288,0.008599999878141616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,3072,0.010024888647927178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,1536,0.08993511067496406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,7168,0.04270577761862013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,4096,0.04145510991414388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,3072,0.007516444557242923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,2048,0.11163822809855144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,128,0.00795377790927887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,6144,0.02938844429122077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,12288,0.19479555553860137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,1024,0.01629599928855896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,6144,0.019261333677503798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,10240,0.023011555274327595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,8192,0.08822755681143866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,3072,0.006855999843941794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,6144,0.6891155772738986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,10240,0.9294604195488824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,4096,0.016809778081046212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,16384,3.567135281032986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,4096,0.022496889034907024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,2048,0.010247999595271217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,65536,0.21305867036183676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,6144,0.10602488782670762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,4096,0.05009955498907301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,3584,0.010016000105275048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,128,0.0027724444452259275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,3072,0.01310755560795466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,64,0.014706666270891825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,3584,0.0070497774415545994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,2560,0.010993777877754636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,128,0.004318222403526306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,2048,0.3821920024024115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,64,0.0046728890803125174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,10240,0.34200533231099445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,128,0.0038071109188927543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,8192,0.0554711123307546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,4096,0.011140444212489657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,5120,0.015646222564909194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,2560,0.00903733323017756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,128,0.030961778428819444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,12288,0.016498666670587327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,65536,0.5177253087361654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,1536,0.005408000200986862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,8192,0.20852711465623644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,8192,0.13598844740125868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,5120,0.013236444857385425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,128,0.004100444416205089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,8192,0.015531554818153381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,1536,0.00980355590581894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,32,0.006350222147173352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,3072,0.09709511200586955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,65536,1.3035359912448459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,32,0.005195555587609609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,16384,0.5496435695224339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,7168,0.019846222466892667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,32,0.012019555601808759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,2560,0.04575466778543261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,8192,0.02835822105407715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,65536,0.05130044288105435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,7168,0.5429137547810873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,65536,3.8895242479112415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,7168,0.01536355581548479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,2048,0.006690666493442323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,1024,0.004611555486917496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,65536,0.25338223245408803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,2560,0.021778666310840186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,64,0.0027066665805048416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,7168,0.011973333027627734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,1536,0.21035555998484293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,6144,0.015460444821251763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,65536,0.3154968950483534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,128,0.009943111075295342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,256,0.003120888852410846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,12288,0.010768888725174798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,4096,0.014684443672498068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,10240,0.019091554813914828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,10240,0.01809955636660258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,3584,0.007976888782448238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,1024,0.006115555349323485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,5120,0.03028622269630432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,1536,0.004679999831649992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,3072,0.011137777732478248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,10240,0.033965332640541926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,1024,0.005394666559166378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,2560,0.06267111168967353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,768,0.005246222019195557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,6144,0.06891200277540419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,512,0.0075911109646161394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,32,0.004972444640265571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,3072,0.019151111443837483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,256,0.007998222278224098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,1536,0.004978666702906291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,5120,0.07295111152860853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,64,0.003378666730390655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,65536,0.07810755570729573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,2560,0.01402844488620758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,2048,0.00544799988468488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,5120,0.010142222046852112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,7168,0.31131378809611004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,10240,0.017094223035706412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,256,0.005151111218664381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,3072,0.05946133534113566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,3072,0.006243555496136348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,65536,0.18339288234710693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,32,0.002871111035346985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,32,0.003478222423129611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,512,0.006268444574541516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,64,0.011390222443474663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,10240,0.029025776518715754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,32,0.007439999944633908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,32,0.014190221826235453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,4096,1.6508248647054036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,5120,0.012106666962305704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,65536,0.07869688669840495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,2560,0.01555288831392924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,5120,0.015318221516079374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,2560,0.01939911146958669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,6144,0.010050666828950247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,2048,0.039705776506000094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,4096,0.010781333678298526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,3072,0.009666666388511658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,10240,0.4448551072014703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,12288,0.05505955550405714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,128,0.007977777885066139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,2048,0.007148444652557373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,10240,0.0373564428753323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,512,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,768,0.005728888842794631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,2560,0.008783111141787635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,32,0.002445333326856295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,4096,0.016490666402710807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,64,0.02240088913175795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,3072,0.021798221601380244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,3584,0.006630222416586346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,6144,0.03684622380468581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,12288,0.015948444604873657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,10240,0.038430223862330117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,3072,0.010229333407349056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,512,0.003216000066863166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,8192,0.04019022319051955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,12288,0.25166577763027614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,512,0.004599110947714912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,7168,0.011039111349317761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,6144,0.008797333472304875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,2048,0.007232888705200619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,7168,0.027131555808915034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,64,0.0046862223082118565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,3584,0.007896000312434303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,65536,0.1392364501953125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,8192,1.0705653296576607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,65536,0.1078924470477634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,512,0.005063999858167437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,3072,0.011311111350854238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,6144,0.06868977679146661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,768,0.0037875556283526947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,7168,0.05117155445946587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,7168,0.04819111029307047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,1024,0.005748444547255834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,6144,0.006396444307433234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,256,0.011347555451922946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,5120,0.011765333513418833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,128,0.0070222218831380205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,1024,0.39845866627163357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,3584,0.16445688406626383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,3584,0.010009778042634329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,10240,0.015013333823945789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,768,0.003576000117593341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,8192,0.0207324441936281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,32,0.00258933338854048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,16384,0.04034844372007582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,5120,0.014468444718254937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,16384,0.020908445119857788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,256,0.00655733338660664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,7168,0.08228888776567247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,12288,0.03760177890459696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,10240,0.011634666886594562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,6144,0.009001777403884465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,1536,0.006549333532651265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,32,0.004913777940803104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,10240,0.12630577882130942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,5120,0.011043555206722684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,3072,0.01569955547650655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,6144,0.03852977686458164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,7168,1.414264890882704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,768,0.06270933151245117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,8192,0.33518399132622617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,64,0.008868444297048781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,768,0.006702222343948152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,8192,0.011384000380833944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,16384,0.12110755178663467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,65536,0.11984266175164117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,2560,0.008264000217119852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,3584,0.007685333490371704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,256,0.004484444442722532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,4096,0.013252444565296173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,1024,0.007918222082985772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,7168,0.029735998974906072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,2560,0.051233778397242226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,16384,0.04157688882615831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,3584,0.006739555547634761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,512,0.005888888819350137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,16384,0.05122311247719658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,16384,0.02428177826934391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,1536,0.01588000026014116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,64,0.0029484445436133277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,7168,0.018975999620225694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,128,0.003999999827808804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,32,0.031397332747777305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,10240,0.011307555768224927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,10240,0.017887110511461895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,1024,0.00904977818330129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,1536,0.013535111314720578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,65536,2.090703116522895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,2560,0.0064693333374129395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,6144,0.5813448694017198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,4096,0.013638222383128272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,7168,0.012265777422322167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,128,0.006755555669466655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,1536,0.009167999857001835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,1024,0.025557332568698462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,2048,0.007677333222495184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,8192,0.014713777436150445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,65536,0.821708467271593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,65536,0.18495910697513154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,7168,0.0181342214345932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,6144,0.0350853337181939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,6144,0.10829244719611274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,10240,0.05097066693835788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,8192,0.031048887305789526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,6144,0.00983911090426975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,8192,0.011604444848166572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,512,0.005397333453098933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,768,0.00870400004916721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,16384,0.19135022163391113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,2048,0.007416889071464539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,6144,0.47575823465983075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,2048,0.0751039981842041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,2560,0.009200000100665623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,3072,0.011063111325105032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,65536,0.06847555769814385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,12288,0.5909635755750868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,512,0.003730666720204883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,3584,0.008928889201747047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,64,0.006581333362393909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,768,0.01297155519326528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,6144,0.20548089345296225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,768,0.005163555757866966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,8192,0.045477333996031016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,2048,0.03378399875428941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,2560,0.009766221874290043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,5120,0.6383751233418783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,256,0.011333333121405708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,1024,0.005024888863166173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,10240,0.03620533479584588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,128,0.0027751111321979095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,10240,0.03445511062939962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,512,0.005471999860472149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,3584,0.3207911120520698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,512,0.0053226666318045715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,1024,0.04185600082079569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,5120,0.038247111770841814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,3072,0.01311555587583118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,1536,0.16259645091162786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,3072,0.005792888916201062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,512,0.0075795559419526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,1024,0.11695733335283066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,2560,0.008544888761308458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,1536,0.6129644181993272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,12288,0.05247733328077528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,768,0.003801777958869934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,10240,0.06662311156590779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,1536,0.006035555568006303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,7168,0.09062844514846802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,16384,0.2759839958614773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,7168,0.006220444622966979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,64,0.003034666594531801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,7168,0.030192000998391047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,2560,0.01184266640080346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,3072,0.014128888646761576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,8192,0.017892445127169292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,16384,0.5481946733262804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,12288,0.3976248900095622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,16384,0.04545066754023234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,5120,0.12345422638787164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,256,0.0042053332759274375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,16384,0.01389244364367591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,4096,0.014305777019924588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,2560,0.006732444382376141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,12288,0.054831110768847995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,8192,0.0214675549003813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,4096,0.016155555844306946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,12288,0.043812443812688194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,2560,0.022414222359657288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,3072,0.02164711058139801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,2560,0.014518222875065274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,10240,0.01754577789041731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,12288,0.00853066643079122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,512,0.005799110978841782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,64,0.002781333401799202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,32,0.0052328887912962176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,65536,0.013232889274756113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,64,0.006576889091067844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,7168,0.2729946772257487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,2048,0.005570666657553778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,768,0.009202666580677032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,3072,0.02072622213098738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,8192,0.013845333622561561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,1024,0.004030222280157937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,3584,0.008127999802430471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,16384,0.12245955732133652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,5120,0.03176800078815884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,65536,0.21925955348544648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,16384,0.029595557186338637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,4096,0.020779555042584736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,1024,0.007937777373525832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,10240,0.1878408855862088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,10240,0.017131555411550734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,1024,0.008809777597586313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,5120,0.011966221862369113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,64,0.0025555555605226094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,256,0.004514666481150521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,2560,0.2020826604631212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,12288,0.04358844293488396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,1024,0.0070159998204973005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,5120,0.030571556753582422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,768,0.06964355707168579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,768,0.0058551111982928375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,64,0.006579555571079254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,3584,0.00962222201956643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,12288,0.020054222808943856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,7168,0.1950746642218696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,1536,0.06015466981463962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,128,0.034527109728919134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,4096,0.011170667078759937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,10240,2.080545849270291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,256,0.0048942222363419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,12288,0.01926399932967292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,7168,0.02985422147644891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,256,0.004953777624501122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,3072,0.008585777547624376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,4096,0.02498400045765771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,5120,0.39730577998691136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,768,0.006164444403515921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,512,0.008394666843944127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,7168,0.6952400207519531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,2560,0.016442666451136272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,1024,0.1320942242940267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,2560,0.012860444684823355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,1024,0.004001777619123459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,2560,0.08349511358473037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,3072,0.009647111097971598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,16384,0.07204800181918673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,1536,0.014144889182514615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,512,0.003650666524966558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,1536,0.02404888967672984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,2560,0.021506667137145996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,7168,0.014540443817774454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,256,0.005250666704442766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,768,0.004991111242108875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,3584,0.017318222257826064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,1024,0.004872888740566042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,65536,0.1250808901256985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,10240,0.048955556419160634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,1024,0.008767111433876885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,2560,0.011051555474599203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,10240,0.025309332542949255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,768,0.0052968888647026485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,2560,0.009316444396972656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,1536,0.004573333180612988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,5120,0.02167555524243249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,12288,0.018920888503392536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,1536,0.006776000062624614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,32,0.006955555743641323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,256,0.00299822228650252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,1024,0.013282666603724161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,2560,0.013784888717863293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,65536,0.02621600031852722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,256,0.006222222414281633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,4096,0.014170666535695394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,8192,0.014226666755146451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,256,0.005214222189452913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,32,0.004648000001907349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,3584,0.006969778074158563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,2048,0.1552320056491428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,6144,0.02401777770784166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,768,0.003552000141806073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,7168,0.030604445272021826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,12288,0.03344710999064975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,32,0.01642666591538323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,2560,0.008848889006508721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,8192,0.039932443035973444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,2048,0.009928888744778102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,128,0.004318222403526306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,512,0.004921777794758479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,1024,0.005920889063013925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,5120,0.02455555564827389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,512,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,6144,0.40323199166191953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,12288,0.05784089035458035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,3072,0.010792888700962067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,1536,0.005758222192525864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,16384,0.03827288746833801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,4096,0.07118755578994751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,768,0.006773333168692059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,3584,0.040864888164732195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,64,0.005388444496525659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,1536,0.005727111051479976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,12288,0.05072089036305746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,64,0.01388533330625958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,1536,0.005976888868543837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,512,0.005961777849329843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,4096,0.01921422282854716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,12288,0.045016888115141124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,768,0.006619555668698416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,2560,0.08065955506430732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,7168,0.021903999977641638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,12288,0.006561777657932705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,16384,0.01854222185081906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,512,0.004996444616052839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,3072,0.04619555672009786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,65536,6.522310045030381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,6144,0.022209778428077698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,64,0.006232888748248418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,5120,0.05993511279424032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,4096,0.01863822175396813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,512,0.006412444429265128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,1024,0.004447999927732679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,8192,0.348453336291843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,768,0.049301332897610135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,768,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,65536,4.07328626844618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,8192,0.020199111766285367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,16384,0.1363404459423489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,128,0.003855111284388436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,7168,0.03017866611480713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,1024,0.00508355556262864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,3072,0.015574221809705099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,768,0.0058248887459437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,256,0.004612444589535396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,6144,0.02512888941499922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,1024,0.005893333504597346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,3072,0.015554666519165039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,3072,0.011759111450778114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,12288,0.014151111245155334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,16384,0.006884444504976273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,8192,0.015838222371207345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,6144,0.02397511071628994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,1536,0.006115555349323485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,16384,0.020074667202101815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,2560,0.007609777980380588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,2560,0.005591111050711737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,128,0.005395555661784277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,1536,0.006666666517655055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,7168,0.01495822270711263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,10240,0.006351999938488007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,12288,0.01641155613793267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,12288,0.015084443820847405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,5120,0.02347466680738661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,3584,0.01804266704453362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,64,0.0069395556218094295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,512,0.003637333297067218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,16384,2.513315624660916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,7168,0.015948444604873657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,768,0.004877333425813251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,12288,0.01939288940694597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,1024,0.0058871110280354815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,7168,0.009191110730171204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,6144,0.04344622294108073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,2048,0.12372622225019668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,10240,0.016777777009540133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,4096,0.009408888717492422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,3072,0.056567112604777016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,16384,0.5814035733540853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,2048,0.049732443359163075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,2560,0.06501155429416232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,4096,0.01679733395576477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,3072,0.021760000122918025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,5120,0.008667555948098501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,8192,0.044271111488342285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,2048,0.010655110908879174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,3584,0.007699555820888943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,1536,0.034480889638264976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,256,0.003069333318206999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,1024,0.005665777872006099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,10240,0.26512355274624294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,128,0.003321777822242843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,65536,0.10871644152535333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,16384,0.11115554968516032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,8192,0.025459556115998164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,3072,0.011626666618718041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,7168,0.04070133301946852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,7168,0.04645244280497233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,128,0.004618666652176115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,768,0.013360888593726687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,512,0.003354666754603386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,4096,0.0746062199274699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,6144,0.012355555262830524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,3072,0.012154666913880242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,12288,0.023756444454193115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,1536,0.01128888875246048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,768,0.003957333250178231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,3072,0.016948444975747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,256,0.005387555393907759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,512,0.003944888710975647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,32,0.005157333281305102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,7168,0.019274666905403137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,1536,0.011305777562989129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,128,0.0026275554878844153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,128,0.004777777940034866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,2048,0.08555377854241265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,2048,0.01034755590889189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,6144,0.006537777682145436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,65536,0.1478382216559516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,768,0.11630845069885254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,16384,0.038592000802357994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,5120,0.020851555797788832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,3072,0.010447111394670276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,512,0.0035991109907627106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,128,0.06054044432110257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,65536,0.5304293102688259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,3584,0.008054222497675154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,512,0.019380443625979953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,512,0.009363555245929295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,3072,0.010472888747851053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,2048,0.03779644436306424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,10240,0.03609866566128201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,128,0.0027431110954946945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,7168,0.013131555583741931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,64,0.0028231110837724474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,5120,0.029189334975348577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,256,0.003154666680428717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,16384,0.020082667469978333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,16384,0.020472000042597454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,2048,0.14890666802724203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,128,0.00546577779783143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,8192,0.022643556197484333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,16384,0.020042666130595736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,4096,0.006616000086069107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,6144,0.3376897705925836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,64,0.003950222084919612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,128,0.003664888855483797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,10240,0.028318223026063707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,6144,0.013458666702111563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,65536,0.18658844629923502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,2048,0.005675555517276128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,4096,0.0188426673412323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,2560,0.009527111219035255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,10240,0.013152889079517789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,10240,0.017464889420403373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,8192,0.012864888542228274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,3584,0.007881777981917063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,1024,0.00573955559068256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,256,0.00815644446346495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,6144,0.03199022346072727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,5120,0.006745777610275481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,65536,1.964120864868164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,8192,0.032403555181291364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,768,0.006294222341643439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,128,0.0027439999911520216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,3072,0.01628800067636702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,16384,0.015229332778188916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,16384,0.13026132848527697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,64,0.0027226667023367355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,8192,0.03403288788265652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,32,0.013186666700575085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,5120,0.17344444327884248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,5120,0.014306666122542487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,768,0.0035911111368073356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,32,0.009296889106432596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,1024,0.14700444539388022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,6144,0.014992000328169929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,3072,0.034640000926123724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,2560,0.010954666468832227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,64,0.002554666664865282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,8192,0.18168177869584826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,3072,0.02566844390498267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,3072,0.013614222407341003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,32,0.0027822222974565295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,12288,0.009462222456932068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,1024,0.004119111018048393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,512,0.0038364442686239877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,3584,0.0074746666683091064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,32,0.0070159998204973005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,65536,0.06786221928066678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,10240,0.1074133316675822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,64,0.002748444469438659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,1536,0.0070293330483966405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,7168,0.03373333480623033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,256,0.004500444564554426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,5120,0.08928444650438096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,65536,0.12407999568515354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,3072,0.006801777829726537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,128,0.01181422256761127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,3072,0.00943288869327969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,10240,0.194252438015408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,8192,0.01237777786122428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,1024,0.009694221946928237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,1024,0.004280888785918553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,7168,0.028394665982988145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,32,0.0026311110705137253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,10240,0.046375112401114575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,2560,0.006978666616810693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,4096,0.00832711077398724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,16384,0.01903733279969957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,3584,0.017342221405771043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,4096,0.13426132996877035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,512,0.0037902221083641052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,1536,0.012492444780137805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,5120,0.02237777743074629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,128,0.003277333246337043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,65536,0.0664657751719157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,768,0.007497777541478475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,1536,0.01018044435315662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,1024,0.018128888474570382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,1536,0.006149333384301927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,8192,0.03282489048110114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,1536,0.006528888725572162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,8192,0.011900444825490316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,12288,0.07925333579381307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,768,0.007267555428875818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,2048,0.00683822234471639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,32,0.006839111033413146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,12288,1.1792578167385526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,12288,0.02257866660753886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,65536,0.706010659535726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,5120,0.010833777487277985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,7168,0.008825778133339351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,5120,0.007557333343558841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,16384,0.06463288598590426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,16384,0.22642932997809517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,4096,0.007753777835104201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,5120,0.0932168894343906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,3584,0.009284444153308868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,8192,0.036537776390711464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,5120,0.018606222338146634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,32,0.009015999734401703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,16384,0.07264888948864408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,256,0.007476444045702617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,16384,0.3443475564320882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,6144,0.02420444455411699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,10240,0.017276444368892245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,1536,0.04946044418546888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,2048,0.015791111522250705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,512,0.016436444388495553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,512,0.004927111168702443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,16384,0.06705333126915826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,3584,0.04077333211898804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,16384,0.04177066683769226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,6144,0.01255644443962309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,768,0.010204444328943888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,256,0.0055440002017551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,8192,0.01955111159218682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,7168,0.0660711129506429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,3584,0.010620444185203977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,3072,0.0210479994614919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,16384,0.019896888070636325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,1536,0.006416888700591193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,1536,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,1536,0.010151999692122141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,8192,0.02265600032276577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,2048,0.06948444578382704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,3584,3.076010598076714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,10240,0.09777600235409206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,32,0.006099555641412735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,1024,0.011551111108726926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,1024,0.010780444575680627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,2560,0.009840000006887648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,6144,0.024586665961477492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,2048,0.014456000592973499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,8192,0.017533333765135873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,5120,0.024127999941507976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,128,0.0027928888383838865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,2560,0.006494222415818109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,256,0.010270222193664974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,12288,0.5835457907782661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,16384,0.06833955314424303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,7168,0.04623644550641378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,16384,0.05850133630964491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,5120,0.03096888793839349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,1536,0.09476177559958564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,128,0.002817777709828483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,6144,0.019215111931165058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,128,0.004381333374314838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,512,0.005576888720194499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,512,0.008825778133339351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,2048,0.0053546664615472155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,65536,0.03418488966094123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,768,0.003597333199448056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,6144,0.01569422251648373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,8192,0.9428088929918078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,65536,0.08741777473025852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,16384,1.5639866722954645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,2560,0.015654222832785714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,16384,0.0890515579117669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,32,0.006009777800904379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,6144,0.048784001006020435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,3584,0.028816001282797918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,7168,0.02718399961789449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,32,0.010503999888896942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,7168,0.01074399964676963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,1024,0.0061919999619325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,1536,0.006346666564544042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,7168,0.5013262430826823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,3584,0.007263999846246507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,8192,0.017901332841979135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,512,0.015069334043396844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,3584,0.007098666495747036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,16384,0.28644177648756236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,7168,0.02068000038464864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,32,0.024767110745112102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,3584,0.010799110763602786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,3584,0.033639109796947904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,1536,0.005869333528810077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,2048,0.006776888751321369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,512,0.0032595555401510666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,12288,1.4155377282036676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,12288,0.010509333676762052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,16384,0.02211644417709774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,1536,0.0063742221229606206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,6144,0.014018666413095264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,2048,0.011544889046086205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,3584,0.028543111350801256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,768,0.004412444515360726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,768,0.09070577886369492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,2560,0.04184622234768338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,4096,0.006599999964237213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,4096,0.013817778064144982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,2048,0.014128888646761576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,1536,0.012378666963842181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,3072,0.009438222481144799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,65536,0.40264089902242023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,12288,0.06199644671546089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,128,0.004909333255555895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,768,0.007229333122571309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,5120,0.015674667225943673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,2048,0.25032711029052734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,1536,0.03357333276006911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,65536,0.029161777761247423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,2560,0.011834666960769229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,512,0.005057777795526716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,16384,0.0236435549126731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,64,0.0031271111220121384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,12288,0.0210853334930208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,16384,0.7286515765719943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,6144,0.01592266725169288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,12288,0.0173680004146364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,256,0.0059057776298787855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,512,0.017892445127169292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,8192,0.14172800381978354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,3072,0.025965332984924316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,16384,0.05372711022694906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,6144,0.01440000037352244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,768,0.003702222059170405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,256,0.0069271110826068455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,3072,0.008650666309727563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,65536,0.16248977184295654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,7168,0.024706666668256123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,10240,0.015096000499195524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,6144,0.11939822302924262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,512,0.008253333469231924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,65536,0.06868088907665677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,16384,0.01536266671286689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,8192,0.02756533357832167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,256,0.004828444371620814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,256,0.0033697777738173804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,65536,0.21557866202460396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,128,0.011918221910794577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,128,0.01112711098459032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,5120,0.15876532925499812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,256,0.004461333155632019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,128,0.0414648883872562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,3584,0.009362666971153682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,12288,0.08226399951510958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,10240,0.04609688785341051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,10240,0.3168604373931885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,64,0.002781333401799202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,4096,0.03363200028737386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,2560,0.011210666762457954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,7168,0.2527804374694824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,7168,0.013530666629473368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,4096,0.008993777963850234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,5120,0.0413973331451416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,10240,0.017109332813156974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,32,0.003137777869900068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,5120,0.020424889193640817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,5120,0.1720186604393853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,65536,0.07012622223960029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,2048,0.0064995553758409284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,3584,0.016996444927321542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,2560,0.010506666368908353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,3072,0.13396088282267252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,512,0.004845333182149463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,32,0.005490666462315454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,512,0.004879999905824661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,10240,0.036171555519104004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,2560,0.01347733371787601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,16384,1.2813137902153862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,768,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,256,0.008603555460770925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,3584,0.7033804257710775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,6144,0.011976000335481433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,2560,0.04814133379194471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,8192,0.0155395550860299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,3584,0.010268444816271463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,2048,0.007872888611422645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,768,0.003621333175235324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,6144,0.015912888778580558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,768,0.0052897776994440295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,12288,0.047594666481018066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,128,0.004736888739797804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,16384,0.0332951115237342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,512,0.004757333132955763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,128,0.006365333166387346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,1024,0.04205066627926297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,4096,0.057759112781948514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,768,0.005680888891220093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,3584,0.01035377797153261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,8192,0.015413333972295126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,12288,0.02491644521554311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,12288,0.08534489075342815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,3584,0.07487822241253324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,10240,0.043656888935301036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,6144,0.024194666081004675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,256,0.006059555543793573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,2048,0.011100444528791638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,6144,0.03015111221207513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,8192,0.015218666858143277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,65536,0.06591288910971747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,64,0.0064026667839950984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,768,0.034132444196277194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,16384,0.02122755514250861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,3584,0.015240000353919135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,256,0.0033280000918441345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,2048,0.0063031112982167145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,65536,0.2628275553385417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,10240,0.08495466576682197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,2560,0.03198755449718899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,8192,0.06097066402435303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,7168,0.12484799491034614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,4096,0.019321777754359774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,64,0.0026791110220882627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,16384,0.014782221780882942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,1024,0.010575999816258749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,12288,0.0410480002562205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,2048,0.008415111237102086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,1536,0.01496355566713545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,1024,0.010375111467308467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,3072,0.010106667048401302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,1024,0.005773333211739858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,32,0.010889777706729041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,5120,0.014109333356221518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,7168,0.014560000763999091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,6144,0.017130666308932833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,8192,0.01612444387541877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,8192,0.023538667294714186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,1024,0.004027555386225383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,1024,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,256,0.0030542222989930045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,16384,0.019176888797018263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,2048,0.007110222346252865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,8192,0.023248889380031165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,1024,0.00490044429898262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,65536,0.13196888234880236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,768,0.004835555536879433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,5120,0.011535111400816174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,5120,0.00905866672595342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,10240,0.03085955646302965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,7168,0.03672177924050225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,1024,0.003906666818592283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,2048,0.006711111300521427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,12288,0.3114062150319417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,7168,0.13868178261650935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,7168,0.010775110787815519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,12288,0.02441155579355028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,16384,0.045132444964514844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,12288,0.07590044207043119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,7168,0.0383200016286638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,3072,0.01960444450378418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,7168,0.03115644388728672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,4096,0.006509333435032103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,1536,0.24619200494554308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,3584,0.011064888702498542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,7168,0.9801386727227105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,1024,0.007032000356250339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,12288,0.030961778428819444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,8192,0.020068445139461093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,7168,0.029231998655531142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,6144,0.009653333160612319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,3584,0.018000000052981906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,512,0.00554577757914861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,8192,0.5617484516567653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,512,0.004944889081848992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,3072,0.01222222215599484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,128,0.007737778127193451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,3072,0.01734133395883772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,12288,0.053411556614769824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,12288,0.011614222493436603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,65536,0.06521599822574191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,1536,0.004801777915822135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,128,0.004024888906213972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,5120,0.008729777402347988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,3584,0.014170666535695394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,7168,0.011552888486120435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,512,0.009465778039561378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,16384,0.037830223639806114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,10240,0.009925333162148794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,65536,0.2076666620042589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,65536,0.08312800195482042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,8192,0.024344889654053584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,64,0.006619555668698416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,65536,0.305607106950548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,6144,0.007129777636792924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,8192,0.013589333328935834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,512,0.011227555572986603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,3584,0.006937777830494775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,16384,0.09301777680714925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,2560,0.007094221810499827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,3584,0.009431111315886179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,4096,0.2801555527581109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,128,0.005482666608360078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,7168,0.010832000109884473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,4096,0.018127111925019156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,1536,0.00488711107108328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,8192,0.6141555574205186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,4096,0.01501066651609209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,5120,0.009358222285906473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,3584,0.00942844483587477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,2560,0.04865511258443197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,32,0.01383733335468504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,2048,0.00535733335547977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,65536,0.2525431050194634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,3584,0.06498310963312785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,7168,0.03346666693687439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,1024,0.008117333054542542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,3072,0.00664444433318244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,2560,0.0069671107663048645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,512,0.0033377777371141645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,16384,0.05565777752134535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,3072,0.036496000157462224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,10240,0.01791111131509145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,10240,0.02804622385236952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,12288,0.10886400275760227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,7168,0.027254222167862788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,64,0.005168000029193031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,10240,0.37287288241916233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,4096,0.010277333358923594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,1024,0.0041271112859249115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,12288,0.08121066623263888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,4096,0.013878222141000958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,1536,0.005622222191757626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,7168,0.006796444455782573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,65536,0.07903822263081868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,16384,0.05790044201744927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,4096,0.007646222081449296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,1024,0.0151235560576121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,128,0.025008888708220586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,256,0.004298666699065102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,1536,0.006268444574541516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,2560,0.012919999659061432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,3584,0.007601777712504069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,1024,0.00565244464410676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,768,0.00564177789621883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,64,0.002556444456179937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,768,0.0064026667839950984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,1536,0.010503999888896942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,512,0.0034942221310403612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,64,0.005261333452330695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,256,0.005422222117582957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,4096,0.02418488926357693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,6144,0.40483734342787003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,512,0.10208621952268813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,8192,0.0317573348681132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,256,0.008543999658690559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,128,0.004618666652176115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,8192,0.051346666283077665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,7168,0.21229867140452066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,128,0.008630221916569604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,128,0.005405333307054307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,65536,0.31824178165859646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,128,0.008328888979223039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,65536,0.12126666969723171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,1024,0.008533333738644918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,5120,0.012884444660610624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,256,0.02683466672897339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,128,0.010982222027248807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,768,0.0036888888312710654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,4096,0.20314133167266846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,6144,0.02536800007025401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,32,0.0047662220895290375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,1024,0.004055110944641961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,512,0.0042053332759274375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,12288,0.0224924451775021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,10240,0.03696622120009528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,128,0.013739555246300168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,5120,0.012739555703269111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,64,0.011815110842386881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,512,0.013179555535316467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,1024,0.014119999276267158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,7168,0.03465244505140517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,5120,0.01426755554146237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,128,0.008293333152929941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,2560,0.006177777631415262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,3072,0.006362666686375936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,4096,0.7873137791951498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,4096,0.02012977831893497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,32,0.0029564443975687027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,32,0.0028106667515304354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,2560,0.01370044466522005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,7168,0.01694222291310628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,3072,0.014789332946141561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,5120,0.01146577795346578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,10240,0.17596622308095297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,32,0.006355555521117316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,512,0.006831999868154526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,5120,0.017651556266678702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,32,0.0029608888758553397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,7168,0.027320888307359483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,3584,0.011567110816637674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,3584,0.017807111144065857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,2560,0.017735110388861764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,4096,0.008416888614495596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,3584,0.011595555477672152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,256,0.010442666709423065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,3584,0.03261244297027588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,6144,0.019853333632151287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,10240,0.7137751049465604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,6144,0.07181155681610107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,12288,0.026625777284304302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,2560,0.07137066788143583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,3584,0.02331022255950504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,1024,0.00906577789121204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,6144,0.024550222688251074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,12288,0.021394666698243883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,3072,0.010257778068383535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,6144,0.019923556182119582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,3584,0.039292444785436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,32,0.05683110819922554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,6144,0.013140444126394061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,3584,0.12437421745724148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,5120,0.008889777792824639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,256,0.008283555507659912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,2048,0.006511111226346757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,2048,0.06301866637335883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,1024,0.003912000192536248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,6144,0.013091555900043912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,3072,0.016121778223249648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,65536,2.442440880669488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,7168,0.13050577375623915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,3072,0.00964355551534229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,10240,0.020111999577946134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,256,0.005254222287072076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,10240,0.03756977783309089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,512,0.005375111268626318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,4096,0.019679110911157396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,6144,0.013583111266295115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,5120,0.05147111084726122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,16384,0.01663466625743442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,10240,0.1519964403576321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,6144,0.013395555317401886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,10240,0.45276445812649196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,7168,0.03172711200184292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,16384,0.5172044436136881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,5120,0.02730577852990892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,2048,0.008292444050312042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,768,0.009542221824328104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,5120,0.03162666824128892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,3584,0.015822221835454304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,1536,0.009348444640636444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,2560,0.00941066692272822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,5120,0.20072533024681938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,1536,0.02104355560408698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,3584,0.012503110700183444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,5120,0.025263999899228413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,128,0.003987555702527364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,6144,0.012206222448084088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,32,0.043271111117468945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,1536,0.004679111142953237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,7168,0.011761777930789523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,7168,0.1767502228418986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,7168,0.026569777064853247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,2560,0.01126844435930252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,6144,0.013968888256284924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,4096,0.008280889027648503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,7168,0.026956443985303242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,128,0.0027475555737813315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,7168,0.027748444014125403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,64,0.00546488869521353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,3072,0.014525334040323893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,8192,0.01937866707642873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,8192,1.618735949198405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,768,0.012612444659074148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,64,0.027060444156328838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,16384,0.11654222011566162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,512,0.00369777778784434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,12288,0.03200888964864943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,2560,0.012323555847009024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,2560,0.013669333524174161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,7168,0.006409777949253718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,128,0.0028284444577164124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,3584,0.11915022797054714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,12288,0.006346666564544042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,1536,0.007330666813585494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,4096,0.05951822466320462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,512,0.004311111238267687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,7168,0.010503110786279043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,1024,0.10670489072799683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,2048,0.03921866748068068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,8192,0.020638222495714825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,7168,0.022291556000709534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,768,0.0061048890153567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,5120,0.02416177756256527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,256,0.00684799998998642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,64,0.005312888986534542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,32,0.005632888939645555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,1024,0.010909332997269101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,5120,0.010602666272057427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,5120,0.041577776273091636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,8192,0.01939911146958669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,1024,0.005793777604897817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,768,0.003592888928121991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,2560,0.0380711125002967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,3072,0.010008888940016428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,6144,0.01862844493654039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,256,0.0034826666944556763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,256,0.005283555636803309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,256,0.002879111096262932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,128,0.0028400001012616684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,10240,0.014025777578353882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,6144,0.014582221706708273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,1024,0.006549333532651265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,5120,0.10075110859341091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,4096,0.019914666811625164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,4096,0.0251528885629442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,4096,0.0801982217364841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,7168,0.028875556257035997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,256,0.00872622181971868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,128,0.003522666792074839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,10240,0.04802133308516609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,16384,0.0299635562631819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,6144,0.026700443691677515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,1536,0.016005333926942613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,1536,0.012329777909649743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,2560,0.010464000205198923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,65536,0.719662242465549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,65536,0.21525422732035318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,5120,0.010474666953086853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,2048,0.006993778049945831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,65536,0.22147111097971597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,128,0.023395554886923894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,10240,0.025997334056430395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,6144,0.10659022463692559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,8192,0.013207111093733044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,256,0.005836444596449534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,8192,0.016364443633291457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,128,0.003047111133734385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,65536,0.08222577969233195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,8192,0.17305421829223633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,16384,0.01422133379512363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,3584,0.007261333366235097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,2048,0.006640888750553131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,1536,0.004752000172932942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,256,0.011509333219793109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,10240,0.0346684455871582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,10240,0.04090222385194566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,1024,0.008220444122950235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,1536,0.006874666859706243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,6144,0.012531555361217923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,3072,0.008100444244013892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,7168,0.04154133465554979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,7168,0.013274666335847644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,2048,0.045309331681993276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,2048,0.005872000008821487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,16384,0.011852444046073489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,128,0.004991111242108875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,256,0.003994666867785984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,65536,0.05968622366587321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,512,0.5305262141757542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,128,0.005148444324731827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,3072,0.025860443711280823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,8192,0.011777777638700275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,5120,0.2479182349310981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,1536,0.020229332976871066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,1024,0.0184871107339859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,12288,0.04758488800790575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,6144,0.02432800001568264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,128,0.0037546666959921518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,8192,0.09893688890669082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,2048,0.011699555648697747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,1536,0.015959110524919298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,8192,0.02250755495495266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,10240,0.045772443215052284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,6144,0.07944977945751615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,5120,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,10240,0.012264889147546558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,2560,0.014667555689811707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,1536,0.007620444728268518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,16384,0.05476977758937412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,1024,0.00610133343272739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,5120,0.014678221609857349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,7168,0.03314666615592109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,1024,0.006061333335108227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,8192,0.011410666836632622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,4096,0.029488888051774766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,2048,0.4097031222449408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,768,0.012982221941153208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,12288,0.016011555989583332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,1024,0.006362666686375936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,256,0.004977777600288391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,256,0.00462311092350218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,12288,0.13658132818010119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,8192,0.025797333982255723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,512,0.005766222046481238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,768,0.004839999808205498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,5120,0.018264888061417472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,65536,0.07742844687567817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,6144,0.020029332902696397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,12288,1.0594221750895183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,3072,0.024223110742039148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,8192,0.02014400064945221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,7168,0.0107022225856781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,4096,0.07092622253629896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,16384,0.023323555787404377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,512,0.01311555587583118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,6144,0.027151111099455092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,1024,0.007374222079912822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,2560,0.009742221898502773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,3072,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,5120,0.009344889058007134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,10240,0.016171556380059984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,7168,0.021755556265513103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,65536,0.022335110439194575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,2560,0.007391999993059371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,3584,0.0074240002367231585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,512,0.005679111099905438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,65536,2.6575262281629777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,1024,0.0355040000544654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,1024,0.0063235556913746735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,2048,0.024314666787783306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,4096,0.006256000035338932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,5120,0.012882666455374824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,32,0.011610666910807291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,2048,0.012126222252845764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,256,0.004938666605287128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,12288,0.04320711228582594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,10240,0.06408088737063937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,12288,0.14651555485195583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,7168,0.13113955656687418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,10240,0.009436444275909001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,32,0.006680000159475539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,32,0.005949333310127258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,64,0.029056890143288508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,64,0.0027751111321979095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,128,0.00463733325401942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,4096,0.016198222835858662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,3072,0.015164444843928019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,4096,0.03366577625274658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,8192,0.03839200072818332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,6144,0.024079999989933435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,256,0.004750222381618288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,2048,0.005326222214433882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,32,0.0027537778433826235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,6144,0.041504889726638794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,6144,0.01014755583471722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,4096,0.01829333272245195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,1536,0.010207110808955299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,7168,0.02611200014750163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,2560,0.007919111185603673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,16384,0.04308266772164238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,256,0.003546666767862108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,8192,0.021545777718226116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,4096,0.02882666720284356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,2048,0.006238222122192383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,1536,0.0063386667105886675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,10240,0.013927110367351107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,256,0.008254222571849823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,512,0.013203555511103736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,2048,0.006036444256703059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,2560,0.011252444651391772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,64,0.002704000100493431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,64,0.002672000063790215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,3584,0.010056888891590966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,6144,0.009673777553770278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,65536,0.08555644750595093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,2560,0.0074026667409473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,2560,0.0069013333155049225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,64,0.031325333648257785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,7168,0.01145600030819575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,10240,0.02645955483118693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,7168,0.012058667010731168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,768,0.012443555725945367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,2560,0.006005333529578314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,1536,0.005868444426192178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,3072,0.0681155522664388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,1024,0.005060444275538127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,256,0.004247111164861255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,3584,0.011854222251309289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,1536,0.005631111148330901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,4096,0.011185777684052786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,65536,0.0857973363664415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,128,0.0029502221279674103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,768,0.004104888687531154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,128,0.003079111170437601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,8192,0.021779555413458083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,8192,0.011759111450778114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,2048,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,10240,0.0396453340848287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,3584,0.042061332199308604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,5120,0.010061333576838175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,128,0.00776533368561003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,5120,0.022553778356975977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,3584,0.02047911120785607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,6144,0.04153600003984239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,2560,0.006072888771692912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,512,0.0057813334796163775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,768,0.0417119993103875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,3584,0.016602666841612924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,1024,0.007643555601437886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,7168,0.01567911108334859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,2560,0.010518222219414182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,1024,0.0163128889269299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,8192,0.01017155581050449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,256,0.004260444392760594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,2048,0.035764445861180626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,16384,0.028151111470328435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,10240,1.5822470982869465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,4096,0.03080444534619649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,2048,0.009139555195967356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,16384,0.023427555958429973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,64,0.0026577777332729767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,128,0.004126222183307012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,3072,0.0321760012043847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,2048,0.017528888252046373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,128,0.00867733359336853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,3072,0.007875555919276344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,3072,0.08336622185177273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,4096,0.00776444458299213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,512,0.01924622224436866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,8192,0.04784711201985677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,10240,0.04920711119969686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,7168,0.01090400003724628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,6144,0.06763911247253418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,256,0.004056888735956616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,128,0.0051893335249688895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,5120,0.037055111593670316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,8192,0.42745333247714573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,7168,0.0466542210843828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,128,0.0058168888919883305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,128,0.0033617777129014335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,3072,0.018023111753993563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,65536,0.07506933477189806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,4096,0.022239999638663396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,2560,0.006276444428496891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,256,0.005333333379692501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,5120,0.00964355551534229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,65536,0.15270845095316568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,10240,0.021495999561415777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,128,0.0053102220926019884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,3584,0.008631111019187504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,2560,0.02924088968171014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,256,0.00563733321097162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,7168,0.011436444189813403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,16384,0.028231110837724473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,1536,0.0064693333374129395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,1536,0.035030222601360746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,3072,0.008945778012275696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,768,0.0184515549076928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,64,0.0052355556852287715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,12288,0.10432622167799209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,64,0.0026275554878844153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,3584,0.027640889088312786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,512,0.021199999584092036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,4096,0.2864053249359131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,256,0.008951110972298516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,3584,0.008001777860853408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,768,0.008837333156002892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,10240,1.7547431521945531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,6144,0.033768001529905535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,256,0.006510222123728857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,2048,0.006344888773229387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,65536,0.23350932863023546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,3072,0.010312889185216691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,64,0.009027555584907532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,10240,0.01670222149954902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,32,0.005325333525737126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,3072,0.014493332968817817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,12288,0.03952177696757846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,10240,0.01774133410718706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,2560,0.01093066649304496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,7168,0.014130666851997375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,3072,0.04071199893951416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,10240,0.013355555633703867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,512,0.006172444257471297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,1024,0.010681777364677854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,1536,0.011493333511882357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,2048,0.019920888874265883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,12288,0.02773244513405694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,5120,0.022636445032225713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,128,0.008734222087595198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,1024,0.00850755555762185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,2048,0.011936888926559024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,5120,0.00962488849957784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,12288,0.05585244629118177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,256,0.0064533332155810455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,6144,0.048560887575149536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,12288,0.06797333558400472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,6144,0.1643431054221259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,1536,0.013628444737858243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,6144,0.10087733136283027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,12288,0.025759999950726826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,16384,0.027976890405019123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,256,0.06682844294442071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,5120,0.03440533412827386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,768,0.005110222018427319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,5120,0.025103110406133864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,256,0.004550222307443619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,1024,0.007844444778230455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,6144,0.022300443715519373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,6144,0.2273777855767144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,768,0.005362666729423735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,1024,0.006529777828190062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,12288,0.019260444574885897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,64,0.002487999904486868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,8192,0.06854933500289917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,7168,0.019919999771647982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,12288,0.03067555692460802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,12288,0.07485244671503703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,128,0.003107555624511507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,2560,0.005970666805903117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,1536,0.005562666803598404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,7168,0.013304000099500021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,2560,0.007787555456161499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,32,0.002603555512097147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,256,0.004819555415047539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,3072,0.015768888923856948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,5120,0.02095199955834283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,512,0.03258666727277968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,12288,0.006984000404675801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,3584,0.02033244404527876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,32,0.23541066381666395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,8192,2.391936832004123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,12288,0.03212533394495646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,256,0.008388444781303406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,64,0.006236444330877728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,1536,0.013023110727469126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,3072,0.0069448889957533936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,8192,0.011755555868148804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,2560,0.0074480002125104266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,65536,1.5238097508748372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,7168,0.017039111918873258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,64,0.004401777767472797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,2560,0.008388444781303406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,256,0.003439999909864532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,256,0.0032951111594835916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,7168,0.054899556769265064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,8192,0.02288977801799774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,7168,0.011739555332395764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,256,0.006824888702895906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,3072,0.015043555034531487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,768,0.010405333505736457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,12288,0.06297689014010958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,3584,0.008093333078755273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,16384,0.06555466519461738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,768,0.027056000298923914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,12288,1.4803537792629664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,6144,0.013286222186353473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,8192,0.2134071058697171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,1536,0.011196444431940714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,32,0.005132444616821078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,3584,0.02790133489502801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,64,0.0029048888633648553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,12288,0.023472888602150813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,256,0.0036346668170558084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,4096,0.016760889026853774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,2048,0.07995911439259847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,512,0.007598222129874759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,3072,0.008460444708665213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,256,0.004879999905824661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,4096,0.05782488981882731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,3072,0.015449777245521545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,1536,0.11737955941094293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,16384,0.06299289067586263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,128,0.007221333682537079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,2560,0.023570666710535686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,512,0.004761777818202972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,5120,0.18853867053985596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,2560,0.007888889147175683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,12288,0.019502222537994385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,16384,0.06727733214696248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,1536,0.019955555597941082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,8192,0.01536622146765391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,8192,0.023535999986860488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,12288,3.60480965508355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,3584,0.040258665879567467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,4096,0.07287733422385322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,2560,0.01051111105415556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,2560,0.00887288898229599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,65536,0.7557039790683322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,65536,0.08589155806435479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,128,0.0028782222006056043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,1536,0.005626666463083691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,2048,0.00720711135201984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,3072,0.010122666756312052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,65536,0.22517689069112143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,65536,0.13935111628638372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,5120,0.009219555391205682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,12288,3.0771314832899304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,3584,0.01715555621518029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,65536,0.06802666849560209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,512,0.005955555372767978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,3072,0.008449777960777283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,2048,0.011520889070298938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,7168,0.01124888906876246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,7168,0.07922577857971191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,768,0.008127999802430471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,7168,0.05204533206091987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,3072,0.01483022173245748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,1024,0.005929777605666055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,12288,0.02999911043379042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,10240,0.02070755594306522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,768,0.004358222087224324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,768,0.018979556030697294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,8192,0.037823110818862915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,1536,0.009570666485362584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,2048,0.04886044396294487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,3584,0.2621928850809733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,65536,0.7725342114766439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,768,0.011145778000354767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,768,0.1187102264828152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,512,0.011554666691356234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,10240,0.01771111124091678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,2048,0.04675288995107015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,8192,0.03931822379430135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,16384,0.12670932875739202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,10240,0.011658666862381829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,32,0.008832889298597971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,8192,0.7552373144361707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,32,0.009191999832789103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,32,0.005767111149099138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,3072,0.02148799929353926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,3584,0.012144889268610211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,2048,0.16220533847808838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,2560,0.013814222481515674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,512,0.007247111035717859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,65536,0.08216622140672472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,512,0.005568888866239124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,8192,0.007927999728255803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,7168,0.012740444805887012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,7168,0.012127111355463663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,7168,0.3854240046607123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,16384,0.057244446542527944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,6144,0.10455466641320123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,2560,0.006626666833957036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,3072,0.011796444654464722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,65536,0.06594933403862847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,3584,0.06829600201712714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,5120,0.03416444526778327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,10240,0.04758222235573662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,1536,0.013567999833159976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,256,0.005394666559166378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,512,0.07477688789367676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,10240,0.6843573252360026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,16384,0.027922666735119287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,5120,0.041173332267337374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,32,0.012484444512261285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,65536,0.2111848857667711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,16384,0.036778665251202054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,128,0.0041991112132867174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,5120,0.6648835606045193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,8192,0.016321778297424316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,512,0.02275200022591485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,3072,0.008104888929261101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,2048,0.39345065752665204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,2048,0.02124622298611535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,256,0.006120000034570694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,256,0.0039013334446483185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,65536,0.14109867148929173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,512,0.003998222036494149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,512,0.005981333139869902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,6144,0.10147111283408271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,128,0.010718222293588849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,64,0.005027555757098728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,1536,0.010651555326249866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,1536,0.0058000000814596815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,512,0.007445333732499017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,128,0.0038364442686239877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,768,0.003690666622585721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,2560,0.012074666718641916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,4096,0.009143999881214565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,1536,0.033909334076775446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,768,0.011244444383515252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,1536,0.0075786668393347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,5120,0.022653333014912073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,128,0.0029519999192820657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,10240,0.04112889038191901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,64,0.004993777722120285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,65536,0.3796204460991754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,3584,0.009232000344329411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,128,0.01163644426398807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,3072,0.006856888946559694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,1024,0.010044444766309526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,1536,0.02717422280046675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,128,0.004055110944641961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,4096,0.008162666526105668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,8192,0.047156443198521934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,512,0.03200622068511115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,256,0.005375999957323074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,4096,0.012476444244384766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,12288,0.034336888127856784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,16384,0.49601157506306964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,5120,0.017671111557218764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,256,0.0030862221287356485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,1024,0.0036888888312710654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,8192,0.03645155496067471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,4096,0.011386666860845355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,7168,0.01812177730931176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,128,0.013439111411571503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,3584,0.02369955513212416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,128,0.0053119998839166425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,12288,0.03657244311438666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,64,0.02366133365366194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,2048,0.005939555664857228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,32,0.0029066666546795103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,16384,0.06048711140950521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,16384,1.3424515194363063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,8192,0.014505777094099255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,16384,0.027664000789324444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,7168,0.04690133200751411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,10240,0.08142577939563327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,128,0.0059075554211934405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,1536,0.016799110505315993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,2048,0.02420266634888119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,1024,0.016576889488432143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,16384,0.0709386666615804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,3072,0.0074515557951397365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,65536,0.06953244739108615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,4096,0.1817244423760308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,4096,0.020358221398459542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,2560,0.00740533322095871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,10240,0.07743289073308308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,3584,0.01943111088540819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,1024,0.035359111097123884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,3072,0.010372444159454769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,256,0.0027928888383838865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,512,0.0074755557709270054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,768,0.004736888739797804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,1536,0.013567999833159976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,256,0.046784000264273755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,8192,0.006808888994985157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,2048,0.011295110815101199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,3584,0.0616497794787089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,4096,0.04905866583188375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,7168,0.13024622864193386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,4096,0.009351111120647853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,2560,0.01496088835928175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,1024,0.008927110996511247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,10240,0.1275413301255968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,1024,0.009300444689061906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,2048,0.02186222208870782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,256,0.006339555399285422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,7168,0.0350213348865509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,512,0.08284533023834229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,4096,0.02050933407412635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,256,0.0029111111329661477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,5120,0.09687822394900852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,12288,0.1073911123805576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,5120,0.014488000008794995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,10240,0.029000888268152874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,16384,0.024917332662476435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,5120,0.009242666264375051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,4096,0.020536889632542927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,5120,0.013032889200581444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,2560,0.051615999804602734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,256,0.003142222141226133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,10240,0.013423110875818463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,3072,0.006859555426571105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,5120,0.019336889187494915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,3072,0.017301332619455125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,4096,0.01202044470442666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,2048,0.009754666851626502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,256,0.005459555735190709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,1024,0.010633777413103314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,2048,0.009602666729026372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,3584,0.02974577744801839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,3584,0.008719999757077958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,768,0.01917866700225406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,128,0.002741333304180039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,512,0.005534222142563925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,2560,0.011156444748242697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,5120,0.22427111201816133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,256,0.010588444769382477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,12288,0.11928889486524795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,7168,0.014256889621416727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,32,0.005056000004212062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,3584,0.010564444793595208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,1536,0.010808889236715106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,1024,0.005057777795526716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,64,0.004574222283230888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,64,0.002842666581273079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,16384,0.04264266623391045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,16384,0.024146666129430134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,3072,0.008089777496125963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,8192,0.03150044547186957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,768,0.0037946667936113146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,16384,0.17274667157067192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,128,0.40386221143934464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,128,0.003962666624122196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,6144,0.026836444934209187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,6144,0.16840178436703154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,8192,0.034372445609834455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,1536,0.011928000383906893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,4096,0.009226666556464301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,10240,0.047373334566752114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,12288,0.08102044132020739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,5120,0.3397040102216933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,512,0.0037919998996787598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,12288,1.1807484096950955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,3584,0.012627555264366997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,12288,0.01257333325015174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,7168,0.0495306650797526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,128,0.009479111267460717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,2048,0.009691555466916826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,5120,0.01352355546421475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,65536,0.07558488845825195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,512,0.004493333399295807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,10240,0.1614417764875624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,2560,0.012359110845459832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,8192,0.03479199939303928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,512,0.007246221933099959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,256,0.0033733333564466904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,128,0.008519111408127679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,16384,0.006586666736337874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,12288,0.07663910918765597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,768,0.02110133402877384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,1024,0.004977777600288391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,16384,2.6646728515625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,4096,0.00775911079512702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,512,0.05330844389067756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,5120,0.07412622372309367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,7168,0.024430221981472437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,2560,0.033837334977255926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,8192,0.09002044465806748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,2560,0.007222221957312689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,768,0.016336000627941556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,3584,0.013820444544156393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,2048,0.007727999654081132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,1024,0.004080888711743885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,65536,0.5346675448947483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,8192,0.01511111193233066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,2048,0.01054844425784217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,1536,0.006636444479227066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,32,0.006913777854707506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,64,0.004436444491147995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,8192,0.034105777740478516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,512,0.007962666451931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,8192,0.017416000366210938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,5120,0.08849155240588719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,768,0.005572444448868434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,64,0.004058666527271271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,256,0.005148444324731827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,32,0.0029804443733559716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,768,0.0057955553962124725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,3072,0.007346666521496243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,5120,0.04002933369742499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,128,0.007919111185603673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,2560,0.021385777327749465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,2560,0.01850755512714386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,2048,0.010866666833559671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,10240,0.04812177684572008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,16384,0.021831999222437542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,2048,0.06540089183383517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,7168,0.02750844425625271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,3072,0.015028445257080926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,4096,0.008461332983440822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,4096,0.611112912495931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,32,0.005299555758635203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,32,0.006154666758245892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,10240,0.02533066603872511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,7168,0.01424800025092231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,1024,0.00545155546731419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,128,0.007679111427730984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,12288,0.0351200004418691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,7168,0.02873777680926853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,16384,0.03484444485770331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,512,0.00674222202764617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,10240,0.021909332937664453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,128,0.005131555514203178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,3584,0.6960106425815158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,2560,0.00958577791849772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,3072,0.01109155515829722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,12288,0.04207377632459005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,8192,0.055903110239240855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,1024,0.008213332957691616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,128,0.004288889053795072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,16384,0.016664889123704698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,2560,0.011593778100278644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,768,0.006724444528420766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,2048,0.014911999305089315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,4096,0.008397333323955536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,5120,0.11298578315311009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,16384,0.07410311036639743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,12288,0.02914488977856106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,2048,0.011925333076053195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,12288,0.07249689102172852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,10240,0.05651466714011299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,2560,0.2268657684326172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,512,0.006254222244024277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,16384,0.04981866810056898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,2560,0.014293332894643148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,65536,0.222396453221639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,7168,0.011042666931947073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,10240,0.04069777660899692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,32,0.002573333266708586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,512,0.0047333331571684945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,2560,0.024431111084090337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,5120,0.04144622219933404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,5120,0.019521777828534443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,32,0.002587555597225825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,768,0.05180533395873176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,16384,0.2807617717319065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,10240,0.01716444392999013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,12288,0.033562666840023465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,7168,0.014099554883109199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,16384,0.11937510967254639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,16384,0.01829777823554145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,1536,0.005594666633341048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,6144,0.08820533090167576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,512,0.01182577759027481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,8192,0.3885644541846381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,128,0.005099555684460534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,16384,0.5290897687276205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,64,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,64,0.0046817776229646476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,4096,0.01571644511487749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,128,0.010812444819344414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,512,0.011137777732478248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,512,0.004041777716742622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,64,0.005503111001518037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,7168,0.49265511830647785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,32,0.00396799999806616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,512,0.04211289021703932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,1536,0.0074577778577804565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,256,0.0058248887459437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,12288,0.015511999527613321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,32,0.0069422221018208405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,3072,0.016752888758977253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,3072,0.02551022171974182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,256,0.004679999831649992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,10240,0.025472000241279602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,512,0.0042408886882993905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,6144,0.022281777527597215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,1024,0.004909333255555895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,4096,0.012011555333932241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,5120,0.009890666438473595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,2048,0.049048887358771436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,512,0.006875555548402999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,16384,0.12188711431291367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,10240,0.03642400105794271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,256,0.007135999699433644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,1536,0.018062222335073683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,768,0.005797333187527127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,4096,0.018037334084510803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,16384,2.4423654344346786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,3072,0.012057777908113269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,7168,0.24824799431694877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,6144,0.016003555721706815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,5120,0.047482666042115956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,512,0.024912888805071514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,65536,0.022634666826989915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,6144,0.028508444627126057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,7168,0.23496267530653211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,1024,0.006275555739800136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,3584,0.011496889094511667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,16384,0.00589155571328269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,65536,0.11294221878051758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,65536,0.0950968861579895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,5120,0.07163288858201769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,768,0.003887999802827835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,10240,0.03105955653720432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,6144,0.017987555927700467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,7168,0.0909786687956916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,1024,0.010046222143703038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,1536,0.009739555418491364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,256,0.003909333298603694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,5120,0.012277333272827996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,2560,0.008592888712882996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,1536,0.008907555705971187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,768,0.008027555214034187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,64,0.005195555587609609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,128,0.004875555634498596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,2048,0.00535999983549118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,1536,0.004328888737493091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,12288,0.03047200043996175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,6144,0.016188444362746347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,5120,0.009668444593747457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,7168,0.12814844979180232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,3584,0.011804444094498953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,2048,0.006642666541867786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,32,0.005140444470776452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,8192,0.07538488838407728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,65536,0.06705333126915826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,8192,0.010053333308961656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,1024,0.006031111296680238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,1536,0.006480000085300869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,64,0.006575999988449945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,7168,0.04721066686842176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,10240,0.1384684509701199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,8192,0.017575111654069688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,512,0.010473777850468954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,65536,0.16003289487626818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,3072,0.009456000394291347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,2560,0.01884355478816562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,12288,0.06652266449398465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,128,0.004007110993067424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,8192,0.01425155500570933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,1024,0.0064399999876817065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,512,0.01423288881778717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,512,0.004857777721352047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,7168,0.013730666703648038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,3584,0.010758221977286868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,2048,0.01240355521440506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,512,0.05094933178689745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,256,0.005446222093370225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,1536,0.0063279999627007385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,32,0.002475555572244856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,4096,0.010831111007266574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,12288,0.25772444407145184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,6144,0.850572427113851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,512,0.015048889650238885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,3584,0.008622222476535374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,8192,0.07183555761973064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,5120,0.0086666668454806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,3072,0.007613333563009898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,2048,0.052038222551345825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,128,0.006591111007663939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,16384,0.07972177531984118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,3584,0.9947271347045898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,7168,0.021596444977654353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,128,0.006210666563775804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,5120,0.011346666349305047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,768,0.009218666288587783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,65536,0.23567289776272246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,32,0.011534222298198275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,32,0.005126222140259213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,256,0.005689777847793367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,4096,0.032256000571780734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,65536,0.26626133918762207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,128,0.004960000101062986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,1024,0.05234044459131029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,2560,0.013794666363133324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,3584,0.007817777494589487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,8192,0.03038755390379164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,1536,0.01239822225438224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,3072,0.006823110911581252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,16384,0.07843733496136136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,6144,0.01591822173860338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,3072,0.020824000239372253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,12288,0.022072888082928125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,10240,0.18418310748206246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,768,0.0036391110883818734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,768,0.008105778031879002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,6144,0.023883556326230366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,2560,0.013576889203654395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,6144,0.04304000073009067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,128,0.007748444047239091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,12288,0.03677333394686381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,512,0.005111999809741974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,5120,0.034584889809290566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,256,0.004606222112973531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,16384,0.022200889057583276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,512,0.0034728890491856467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,1024,0.005564444594913059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,8192,0.5799386766221788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,4096,0.012502222425407834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,7168,0.05590755409664578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,2048,0.15239289071824816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,10240,0.04313066601753235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,16384,0.018886221779717337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,768,0.006583999842405319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,2560,0.018883556127548218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,16384,0.0749173363049825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,10240,0.021114667256673176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,5120,0.009495110975371467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,3072,0.007289778027269576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,16384,0.01814844376511044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,3584,0.012813333008024426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,5120,0.014863999353514777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,256,0.005192000004980299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,3584,0.027457776996824477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,1536,0.03713422351413303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,8192,0.06578044758902656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,7168,0.029641777276992798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,3072,0.007309333317809635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,1536,0.017151110702090792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,1536,0.012063110868136087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,2048,0.008345777789751688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,768,0.006027555714050929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,1024,0.006078222145636876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,5120,0.022295999858114455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,1536,0.004601777841647466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,4096,0.10475555393430923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,512,0.025076443950335186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,2048,0.14883644051022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,65536,0.8259279992845324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,256,0.005018666800525454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,768,0.0037395556767781577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,16384,0.10257866647508408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,1536,0.008750222623348236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,10240,0.006802666518423293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,10240,0.014934221903483072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,12288,0.13622132937113443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,10240,0.039606223503748574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,32,0.0068959999415609576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,65536,0.10172355837292141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,1024,0.011884444289737277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,2048,0.014498665928840637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,512,0.0047155556579430895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,5120,0.008438222110271454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,512,0.005439110928111606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,10240,0.009697777529557547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,512,0.216231107711792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,4096,0.043417778280046254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,12288,0.03937066594759623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,1024,0.014925332532988654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,3072,0.015787555111779105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,3584,0.00756888919406467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,6144,0.011235555840863122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,65536,0.5037795702616373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,7168,0.021048888564109802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,1536,0.004591111093759537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,5120,0.11682844161987305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,768,0.004440000073777305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,3584,0.12517333030700684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,2048,0.009904888768990835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,1536,0.006657777975002925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,4096,0.0075697774688402815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,4096,0.008264888491895463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,128,0.002757333426011933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,3584,0.006802666518423293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,64,0.005464000006516774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,12288,0.03143466512362162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,16384,0.025634666283925373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,16384,0.046931554873784385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,65536,0.039431111680136785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,7168,0.06935733556747437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,16384,0.07781244648827447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,128,0.04273866613705953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,10240,0.03934311204486423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,6144,0.1142088837093777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,128,0.00328533330725299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,4096,0.01126488877667321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,6144,1.028630256652832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,512,0.025771554973390367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,32,0.003593777616818746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,768,0.04211911227968004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,768,0.013741333451535968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,8192,0.13736177815331355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,12288,0.022648889157507155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,256,0.0038515557017591265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,128,0.006043555421961679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,6144,0.010413332945770687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,3584,0.07901510927412245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,128,0.005176888985766305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,1536,0.007896000312434303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,3584,0.010833777487277985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,6144,0.04690577917628818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,16384,0.025032000409232244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,8192,0.015399111641777886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,768,0.00408000002304713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,512,0.005314666777849197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,1024,0.05230577786763509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,65536,0.019031110737058852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,65536,0.07020977470609877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,16384,0.006252444452709622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,3584,0.009413333402739631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,32,0.03320088982582092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,768,0.029940442906485662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,64,0.0027066665805048416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,256,0.018295110927687753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,2048,0.03864177730348375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,12288,0.02384977704948849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,7168,0.03961599866549174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,1536,0.013340444200568728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,65536,0.14917511410183376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,256,0.0036951113078329298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,2048,0.010080888867378235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,65536,0.14453689257303873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,768,0.037695109844207764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,16384,0.019578667150603402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,32,0.0033128888656695685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,3072,0.027317333552572463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,1536,0.006263111200597551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,768,0.06783999999364217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,12288,0.015188443991872998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,5120,0.0069368887278768755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,2048,0.019328888919618394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,7168,0.010382222632567087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,3072,0.0063546668324205614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,1536,0.00847111145655314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,768,0.0033377777371141645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,6144,0.5397457546657985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,12288,0.015326221783955893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,6144,0.23449156019422743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,2048,0.023431999815834895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,2048,0.007394666473070781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,256,0.005555555638339784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,2048,0.014853333433469137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,32,0.0026942222482628296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,1024,0.005515555540720622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,1024,0.004910222358173794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,4096,0.010608889162540436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,1536,0.03577333357599046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,12288,0.01521511044767168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,8192,0.011459555890825061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,128,0.00546488869521353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,3072,0.0075786668393347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,4096,0.019440000255902607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,1024,0.010257778068383535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,1024,0.006601777755551868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,2048,0.0076453329788313965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,10240,0.014303111367755465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,128,0.0052879999081293745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,65536,0.07342400153477986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,1024,0.005264889034960005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,12288,0.015800888339678448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,256,0.003116444374124209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,3072,0.013094222380055321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,3584,0.009910222556855943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,7168,0.05887822310129801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,6144,0.19795733027988008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,65536,0.1842142211066352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,5120,0.020738666256268818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,1536,0.006264888991912206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,7168,0.02708622150950962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,32,0.002597333242495855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,6144,0.008632000121805403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,4096,0.008184889124499427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,10240,0.0385484430525038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,1536,0.0046160001721647055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,2048,0.05801244576772054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,256,0.006226666685607698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,10240,0.01607644392384423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,768,0.6052284240722656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,2048,0.007537778053018782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,2560,0.0675813357035319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,128,0.0032586666444937387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,6144,0.011333333121405708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,16384,0.12476800547705756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,64,0.005569777968857024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,5120,0.009696889254781935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,65536,0.06709066364500257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,32,0.09377511342366536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,64,0.00979733301533593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,3584,0.1013866662979126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,16384,0.3412675592634413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,512,0.0099982221921285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,4096,0.009383999639087254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,8192,0.015480000111791821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,6144,0.281023104985555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,32,0.0027191111197074256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,8192,0.015375110838148328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,7168,0.00903733323017756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,3072,0.024714666936132643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,7168,0.023896000451511804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,768,0.009525333013799455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,5120,0.08878933058844672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,12288,0.01993333299954732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,16384,0.18965866830613878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,512,0.0033191111352708605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,7168,0.028720888826582167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,10240,0.014560000763999091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,3072,0.02499288817246755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,3584,0.011768889096048145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,2560,0.022838221655951604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,1024,0.008519111408127679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,16384,0.04152889053026835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,65536,0.0749795569313897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,128,0.004695999953481886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,12288,0.08964533276028103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,256,0.005223111146026188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,5120,0.009325332939624786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,32,0.002704888789190186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,16384,0.13403822316063776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,768,0.012614222036467658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,2048,0.006035555568006303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,3584,0.17928889062669542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,1536,0.009666666388511658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,64,0.006736889067623351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,3584,0.015927111109097798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,768,0.004629333400064045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,128,0.0037662221325768363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,1024,0.005551110953092575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,1024,0.014183999763594734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,1536,0.0047182221379545005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,768,0.007888000044557784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,64,0.0031786666562159858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,1024,0.0052782222628593445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,2048,0.015638222297032673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,256,0.006968888971540663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,3584,0.01940088967482249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,5120,0.02072533302836948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,64,0.005294222384691238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,1024,0.019839111301634047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,65536,0.06553244590759277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,16384,0.05179377728038364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,768,0.003670222229427762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,3584,0.008168000313970778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,512,0.005426666802830166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,1024,1.711947547064887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,10240,0.02016088863213857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,1536,0.008431110945012834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,1536,0.02092355489730835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,256,0.004300444490379757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,2560,0.01590399940808614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,3072,0.012791111237472959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,512,0.0033297776761982176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,3584,0.01845066746075948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,768,0.0047182221379545005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,10240,0.02480711042881012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,512,0.005335999859703912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,768,0.005764444255166584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,8192,0.04982666505707634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,3584,0.02330222229162852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,256,0.030533333619435627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,5120,0.006784889019197888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,12288,0.01942399972014957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,10240,1.184523582458496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,10240,0.21754222446017793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,1536,0.010416888528399997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,768,0.006779555645253923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,128,0.005148444324731827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,32,0.007430222299363878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,6144,0.06652888986799452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,8192,0.015286222100257874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,16384,0.03853777713245816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,64,0.0035075553589397003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,1536,0.006481777876615524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,2048,0.006805333412355847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,65536,0.0670800010363261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,12288,0.020264888803164165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,2560,0.09520444605085586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,2560,0.009380444056457942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,768,0.015103111664454142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,10240,0.01865600049495697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,768,0.005598222215970357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,256,0.00490399988161193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,16384,0.046747555335362755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,5120,0.01034400032626258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,512,0.004287999951177173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,4096,0.008360889222886827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,4096,0.018395556343926322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,768,0.009679111341635386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,6144,0.18888000647226968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,2048,0.020658666888872784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,8192,0.06622755527496338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,16384,0.02583822276857164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,128,0.0028595555987623003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,3584,0.04220889012018839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,12288,0.0529075562953949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,768,0.028135998381508723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,2048,0.006147555592987273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,256,0.004288889053795072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,1536,0.055620445145501025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,7168,0.06422399812274508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,768,0.020079111059506733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,6144,0.011896889242861005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,512,0.003522666792074839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,512,0.004675555560323927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,256,0.0028284444577164124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,6144,0.010942222343550788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,1536,0.010136889086829292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,4096,0.016757332616382174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,3072,0.007356444166766272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,2560,0.006858666737874349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,8192,0.01163999984661738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,5120,0.017874666386180453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,6144,0.02050933407412635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,12288,0.04022044605678982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,4096,0.010808000134097205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,6144,0.0075253330998950535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,6144,0.7096693250868055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,256,0.010829333629873065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,12288,0.024506666594081458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,8192,0.06463199853897095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,2560,0.011339555184046427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,16384,0.042675554752349854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,16384,0.24949955940246582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,16384,0.2966889010535346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,7168,0.033542222446865506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,4096,0.014537778165605334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,256,0.011191111471917896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,2560,0.012986666626400419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,512,0.012551999754375882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,3584,0.20549510584937203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,1536,0.005059555586841372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,768,0.015133332875039844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,32,0.005114666703674528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,2560,0.018369777335060965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,65536,0.20868355698055693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,6144,0.009524444739023844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,3072,0.05394133263164096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,65536,0.06606133116616143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,10240,0.016306666864289176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,32,0.004470222112205294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,32,0.00499555551343494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,10240,0.017454221844673157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,2560,0.005973333285914526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,2560,0.007326222128338284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,128,0.00654755574133661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,6144,0.025660443637106154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,8192,0.02566577825281355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,12288,0.042837331692377724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,32,0.006341333190600078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,4096,0.028674665424558852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,2560,0.007782222496138678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,1024,0.006696888970004187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,128,0.003104888937539525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,7168,0.015095111396577625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,7168,0.31937511761983234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,12288,0.46363290150960285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,16384,0.07356266842948066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,8192,0.031113776895735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,7168,0.06897333595487806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,256,0.06606755654017131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,128,0.0038293335172865125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,768,0.0063866666621632045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,64,0.006040888941950268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,7168,0.013967111706733704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,3072,0.01906222270594703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,1536,0.014138667119873894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,1024,0.026387555731667414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,768,0.014927110738224454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,7168,0.01607822212908003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,2048,0.41962578561570907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,12288,0.035896887381871544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,6144,0.01976977785428365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,256,0.003026666740576426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,6144,0.010339555641015371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,256,0.005299555758635203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,1536,0.007983110845088959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,512,0.045138667027155556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,6144,0.01422044469250573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,256,0.004541333350870344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,3584,0.015557333827018738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,512,0.004885333279768626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,8192,0.03435555431577895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,1024,0.005457777943876054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,16384,0.04956266615125868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,10240,0.045815110206604004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,768,0.004325333154863781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,4096,0.014666666587193808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,512,0.0035297779573334586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,256,0.0032764443506797156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,1024,0.005816000203291575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,256,0.004537777768241035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,32,0.0028071111689011254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,6144,0.022143110632896423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,1024,0.012404444317022959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,65536,0.05777421924802992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,5120,0.02297244469324748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,3072,0.010015111002657149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,2048,0.025760889053344727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,32,0.027082666754722595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,4096,0.013438222308953604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,16384,0.024684444069862366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,64,0.0032302221904198327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,4096,0.011254222028785281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,16384,0.009113777842786577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,768,0.005438222239414851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,3072,0.008445333275530074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,3072,0.024509333901935156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,64,0.003474666840500302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,512,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,10240,0.05140533381038242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,2560,0.007236444287829929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,512,0.0032213332338465583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,4096,0.019174221489164565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,768,0.003710222327046924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,4096,0.010185778141021729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,2048,0.006735111276308696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,128,0.0027093332674768236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,3584,0.008582221964995066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,256,0.04680000080002678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,256,0.004834666848182678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,64,0.0034711110509104202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,32,0.0053377776510185665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,256,0.003899555653333664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,8192,0.0063457778758472866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,32,0.006487999939256244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,512,0.005400889035728242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,128,0.00647555540005366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,7168,0.029888000753190782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,65536,0.553741349114312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,1536,0.029072887367672388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,4096,0.07853511306974623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,2560,0.018273777431911893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,65536,0.06363111072116427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,5120,0.012727999852763282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,7168,0.014937778313954672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,16384,0.3215102089775933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,10240,0.054245332876841225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,64,0.0028151110228565005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,65536,0.17289066314697266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,32,0.0047084444926844705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,512,0.007511110769377814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,10240,0.16353332996368408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,65536,3.084443622165256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,10240,0.04653777678807577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,6144,0.01566488875283135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,128,0.004139555411206352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,64,0.0028373334142896864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,4096,0.01276533305644989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,512,0.005793777604897817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,8192,0.030760000149408977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,8192,0.009161777794361115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,16384,0.03801155421468947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,16384,0.049792889091703624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,10240,0.020077332854270935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,6144,0.05380355649524265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,5120,0.013800889253616333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,12288,0.0188044442070855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,8192,0.020625778370433383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,2048,0.018943111101786297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,12288,0.0269777774810791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,3072,0.017309332887331646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,2560,0.007657777931955125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,12288,0.04294755392604404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,2048,0.00665866666369968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,65536,0.29748445087009007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,65536,0.2031724452972412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,10240,0.20075378153059217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,768,0.008041777544551426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,128,0.0210479994614919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,768,0.004062222109900581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,2560,0.015169777803950839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,2048,0.008382221890820397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,4096,0.025608888930744592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,16384,0.037775112522972956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,128,0.007705777883529663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,5120,0.01111822244193819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,8192,0.7591439882914225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,64,0.0033440000067154565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,5120,4.17391120062934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,1536,0.017892445127169292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,128,0.003547555456558863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,5120,0.0221395558781094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,1024,0.005240889059172736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,2048,0.046317332320743136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,7168,0.12616799937354192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,256,0.003051555612021022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,3072,0.015536889433860779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,1536,0.014780445231331719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,768,0.005112888912359874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,1024,0.023408000667889912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,1024,0.0900506642129686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,5120,0.015971556305885315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,10240,0.17235822147793242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,3584,0.06507822540071276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,3584,0.020587555236286588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,2560,0.14671022362179226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,12288,0.06796089145872328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,5120,2.0709465874565973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,32,0.005330666485759947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,8192,0.08646933237711589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,4096,0.01606399979856279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,8192,0.013373333546850415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,65536,0.23583110173543295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,10240,0.018990221950742934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,32,0.00295733329322603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,7168,0.02664533257484436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,3584,0.01198933356338077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,1024,0.006450666735569636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,10240,0.01809599995613098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,256,0.00619555554456181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,6144,0.011163555913501315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,6144,0.010404444403118556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,1024,0.005680888891220093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,7168,0.6261404355367025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,4096,0.011448000040319232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,8192,0.04363822274737888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,1024,0.009720000127951304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,256,0.00434400017062823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,3072,0.016208888755904306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,8192,0.27435021930270725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,512,0.07224622037675646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,16384,0.05624800258212619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,5120,0.012708444562223224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,2048,0.005906666732496685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,10240,0.2772666613260905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,5120,0.012879999975363413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,256,0.0035866668654812705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,6144,0.021459556288189355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,3072,0.009489778015348647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,16384,0.02049511174360911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,8192,0.06260355313618977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,128,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,256,0.008577778107590145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,12288,0.32484179072909886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,3584,0.010168888502650792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,512,0.005102222164471944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,10240,0.033562666840023465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,65536,0.37781866391499835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,10240,0.021196444829305012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,1536,0.016930666234758165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,1024,0.01127911110719045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,4096,0.006527110934257507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,1536,0.00888000014755461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,2048,0.0063493334584765965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,64,0.005125333451562458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,65536,0.5170453389485677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,3584,0.010589333044158088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,6144,0.013152889079517789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,16384,0.04130399889416165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,12288,0.010079111489984725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,64,0.02495199938615163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,2048,0.011007111105653973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,768,0.0047600000268883174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,16384,0.08681777450773452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,768,0.03126488791571723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,768,0.0037715555065208008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,6144,0.0119982221060329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,256,0.005387555393907759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,4096,0.0096124443742964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,2048,0.019896888070636325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,6144,0.046408888366487294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,8192,0.011615110768212212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,7168,0.013815999858909182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,3584,0.013896889156765409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,32,0.016435555285877652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,32,0.041526221566730075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,2048,0.005365333209435145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,10240,0.073926223648919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,5120,0.01107377807299296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,10240,0.07666844129562378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,7168,0.015085332923465304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,10240,0.062031997574700244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,1536,0.008541333178679148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,6144,0.010229333407349056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,256,0.00286666676402092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,10240,0.013571555415789286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,12288,0.017417778571446735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,8192,0.018965333700180054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,768,0.011680888632933298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,6144,0.010048000348938836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,7168,0.014006222287813822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,64,0.006395555618736479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,65536,0.02316355539692773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,128,0.004193777839342753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,4096,0.010464888479974536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,12288,0.20901866753896078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,6144,0.010094222095277574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,7168,0.0143857780430052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,65536,0.22312532530890572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,4096,0.010562666588359408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,8192,0.1141422192255656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,512,0.011977777712874942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,1536,0.009684444301658207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,16384,0.020812445216708712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,2560,0.007093333535724216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,7168,0.03252355588807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,7168,0.01220533334546619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,12288,0.02462755474779341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,768,0.005443555613358815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,10240,0.06398844718933105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,512,0.022429333792792425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,5120,0.012147555748621622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,7168,0.03406577640109592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,512,0.0032497776879204642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,8192,0.012534221841229333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,4096,0.018279999494552612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,2560,0.014760888285107084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,32,0.005630222045713001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,12288,0.03139288889037238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,5120,0.008218666745556725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,64,0.009842666486899057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,4096,0.016763556334707472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,65536,0.03073244293530782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,2560,0.00870311094654931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,5120,0.02985422147644891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,65536,0.01127199994193183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,3072,0.008030222521887885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,2560,0.006511111226346757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,2048,0.07195199860466851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,6144,0.016607999801635742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,6144,0.018984888990720112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,256,0.0029253332565228143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,7168,0.08674844768312241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,128,0.004146666576464971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,8192,0.12212177117665608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,10240,0.013384888569513956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,3584,0.007459555235173967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,5120,0.0067528887755341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,2048,0.0476764440536499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,8192,0.08051466941833496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,2048,0.05468444360627068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,4096,0.07650133636262682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,64,0.0026542221506436667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,3072,0.01586844523747762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,128,0.003585777762863371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,6144,0.03695999913745456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,256,0.0046942221621672315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,7168,0.009222221871217092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,3072,0.010986666712496014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,3072,0.01586666703224182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,32,0.005013333426581489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,3072,0.2038195530573527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,10240,0.022447110878096685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,7168,0.014244443840450711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,4096,0.012847999731699625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,128,0.0042408886882993905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,7168,0.030551999807357788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,2560,0.014004444082578024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,10240,0.044940445158216685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,32,0.00554044461912579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,3072,0.010944888823562197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,1024,0.008058666355080074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,12288,0.020280889338917203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,8192,0.011663999822404651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,512,0.007079111205206976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,2048,0.006536888993448681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,6144,0.02676888969209459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,12288,0.19291910860273573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,128,0.004566222015354368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,3584,0.020407110452651978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,65536,0.0663964417245653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,3584,0.04310133390956455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,2560,0.03740266627735562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,4096,0.02068088948726654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,2560,0.01145066652033064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,3072,0.01845066746075948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,768,0.010116444693671333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,3584,0.020516445239384968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,5120,0.01090844472249349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,256,0.0038906666967603895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,12288,0.21657333109113905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,1536,0.00535111129283905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,1024,0.012348444097571902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,10240,0.9562755160861545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,3072,0.06443022357092963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,1024,0.004576000074545543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,3072,0.015054222610261707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,12288,0.2759182188245985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,3072,0.012433778080675336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,16384,0.031046221653620403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,8192,0.03400977783732944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,5120,0.024442666106753882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,2048,0.019935111204783123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,3072,0.013902222116788229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,16384,0.7878933482699924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,4096,0.037447998921076454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,16384,0.030905776553683813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,12288,0.08333510822719997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,10240,0.027012444204754297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,512,0.006393777827421824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,1536,0.005959110955397288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,64,0.00254488881263468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,10240,0.13317067093319362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,3584,0.01998488936159346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,1536,0.01608888970481025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,4096,0.013535111314720578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,1024,0.01700799994998508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,64,0.006174222048785951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,2048,0.007086222370465596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,7168,0.03503200080659654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,1024,0.004671111288997862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,5120,0.008621333373917473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,65536,0.8079315291510688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,5120,0.060172445244259305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,32,0.006313777632183499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,16384,0.03447377681732178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,2048,0.007228444847795699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,1024,0.01516355574131012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,2560,0.009137777818573846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,3072,0.010625777973069085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,3072,0.011124444504578909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,512,0.007800000409285228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,2560,0.012111110819710625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,64,0.0051893335249688895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,6144,0.005859555469618903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,32,0.004271999829345279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,6144,0.013080889152155982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,16384,0.07543733384874132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,65536,0.08606044451395671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,2048,0.009650666680600908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,1536,0.004831999954250124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,768,0.004769777672158347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,16384,1.6079893112182617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,8192,0.018668444620238412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,512,0.007840889195601145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,10240,0.07476622528500028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,12288,0.01535377734237247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,16384,0.14469333489735922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,16384,0.03914311197068956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,512,0.005196444276306364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,2048,0.006268444574541516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,8192,0.06618222263124254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,32,0.0075048887067370946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,10240,0.028860443168216284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,4096,0.019824888971116807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,65536,0.24510934617784289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,1024,0.005128889034191768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,16384,0.5475529034932455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,256,0.019075555933846366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,4096,0.012726222475369772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,2560,0.006805333412355847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,16384,0.024088889360427856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,6144,0.01850311126973894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,8192,0.019023999571800232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,4096,0.010024888647927178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,10240,0.027678221464157104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,65536,0.06782311201095581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,65536,13.948194715711805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,12288,0.23979200257195366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,1536,0.005631999837027655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,32,0.005129777722888523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,3072,0.021591110361946955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,32,0.0024844445288181305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,4096,0.01684266659948561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,1536,0.00656444455186526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,1536,0.016223111086421542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,1536,0.011804444094498953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,4096,0.009461333354314169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,7168,0.045732445187038846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,512,0.008421333299742805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,3584,0.12343200047810872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,32,0.00814133303032981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,6144,0.026201777988009985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,4096,0.018364444375038147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,2560,0.02017422186003791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,512,0.023924445112546284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,3584,0.009327111144860586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,32,0.0030177777840031516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,4096,0.018925334016482036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,65536,0.1223262283537123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,5120,0.009496000077989366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,3072,0.19211555851830375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,768,0.0052533331844541765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,128,0.00847822262181176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,6144,0.010409778190983666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,3584,0.017419555120997958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,7168,0.06835999753740099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,2048,0.010934222075674268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,1536,0.010008888940016428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,512,0.00794933322403166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,4096,0.13348533047570124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,5120,0.011722666521867117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,12288,0.021852445271280076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,128,0.0042462220622433555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,4096,0.007725333174069722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,16384,0.02237777743074629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,4096,0.012177777787049612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,4096,0.012607110871209038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,128,0.0059057776298787855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,1536,0.004604444321658877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,12288,0.05307288964589437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,4096,0.008548444343937768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,4096,0.043284446001052856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,12288,0.1577733357747396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,16384,0.03363200028737386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,10240,0.01457244488928053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,8192,0.07273955477608575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,2560,0.006025777922736273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,2048,0.01293333371480306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,5120,0.010464888479974536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,6144,0.010148444109492833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,7168,0.09776000181833903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,4096,0.018271999226676095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,7168,0.006364444477690591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,768,0.008204444415039485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,256,0.012978666358523898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,128,0.004313777718279096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,256,0.005919999960396025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,2048,0.016551110479566786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,3584,0.06931733422809176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,7168,0.05489600035879347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,3584,0.18025510840945771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,1024,0.0037573331760035623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,7168,0.04456444581349691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,2560,0.015523556205961438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,6144,0.008398222426573435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,32,0.0026231110095977783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,6144,0.041336887412601046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,256,0.006782222125265334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,256,0.005929777605666055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,3584,0.1129208935631646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,5120,0.02055644492308299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,768,0.0058311112225055695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,5120,0.018239110708236694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,512,0.011539555258221097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,1536,0.011100444528791638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,1024,0.0052764444715446895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,1536,0.00928177767329746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,768,0.004691555682155821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,128,0.0069226668112807805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,65536,0.07282666365305583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,256,0.004619555754794015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,1536,0.012072888513406118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,1024,0.008961777720186446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,2048,0.009087111386987898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,64,0.002921777880854077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,2048,0.006836444553401735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,8192,0.02971733278698391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,2560,0.007080888582600488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,6144,0.2092426617940267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,4096,0.008959111240175035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,16384,0.06254666381412081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,768,0.006190222170617845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,8192,0.053006221850713096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,128,0.009226666556464301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,4096,0.013099555340078143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,128,0.00369422220521503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,10240,0.10733955436282688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,65536,0.11203911569383408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,4096,0.026130666335423786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,2560,0.009915555516878763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,3072,0.008823999928103553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,65536,0.17758577399783662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,2048,0.006274666637182236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,768,0.003522666792074839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,5120,0.01090222183201048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,3072,0.01199288914601008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,128,0.005244444641802046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,5120,0.021831111775504217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,4096,0.0262071109480328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,10240,0.028044445647133723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,1536,0.006966222491529252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,768,0.00776711106300354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,8192,0.027543110979927912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,10240,0.15056355794270834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,512,0.006041777630647023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,128,0.008244444098737504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,16384,0.07166666454739042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,32,0.0028391112056043413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,16384,0.01868088874551985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,2048,0.014135110709402295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,256,0.005162666655249066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,4096,0.02529955572552151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,12288,0.04552622305022346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,3584,0.0185217774576611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,7168,0.01830044388771057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,2560,0.1307448943456014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,16384,0.018217777212460835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,16384,0.02106933295726776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,256,0.008303999900817871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,32,0.007483555210961237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,2048,0.010957332948843638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,8192,0.01143200033240848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,64,0.031552000178231135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,32,0.003118222165438864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,8192,0.05639288822809855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,2560,0.007630222373538547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,32,0.00656355544924736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,512,0.005944889038801193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,16384,0.5381884574890137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,3584,0.012140444583363004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,12288,0.3786551157633464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,16384,0.20512710677252877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,10240,0.017910222212473553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,4096,0.019860444797409903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,1024,0.05133688780996534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,7168,0.34555376900566953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,3584,0.044250667095184326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,5120,0.18193688657548693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,2560,0.0063493334584765965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,10240,0.01789066692193349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,65536,4.914102342393663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,16384,0.030272000365787085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,64,0.005402666827042897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,5120,0.012560000022252401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,7168,0.3377999994489882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,3072,0.4884764353434245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,4096,0.006314666734801398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,64,0.005177777674463061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,6144,0.02810844447877672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,32,0.007012444237867991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,5120,0.016336888074874878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,3072,0.01274400038851632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,2560,0.013372444444232516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,3072,0.008201777935028076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,10240,0.035261332988739014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,128,0.004209777961174647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,5120,0.012512000070677863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,65536,0.033055110110176936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,6144,0.9292657640245227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,7168,0.027259555127885606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,8192,0.025249777568711176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,2048,0.009311111436949836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,3072,0.008937777744399177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,7168,0.014746665954589844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,256,0.005184000151024925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,65536,0.00925422211488088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,512,0.02625600000222524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,768,0.007391999993059371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,8192,0.013029333617952136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,10240,0.035098665290408664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,128,0.005070222334729301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,2560,0.012269333004951477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,3072,0.006890666567616993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,512,0.005173333403136996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,3584,0.010847110715177325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,1536,0.005411555369695027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,4096,0.08844711383183797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,2560,0.008047999607192146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,128,0.0047315553658538395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,512,0.0038480001191298165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,12288,0.006693333387374878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,256,0.0069475554757648045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,256,0.0063946665161185795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,768,0.005912888795137405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,128,0.009343110852771336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,1024,0.00998755544424057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,6144,0.025456888808144465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,512,0.004791111167934206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,512,0.003432000055909157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,6144,0.020954666866196528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,2560,0.6005697780185276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,5120,0.023271110322740342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,1024,0.005511111269394557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,3584,0.012537777423858643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,256,0.0046453335218959385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,128,0.004036444342798657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,12288,0.03331111205948724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,12288,0.5364693535698785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,7168,0.10021244155036078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,32,0.005151999907361137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,768,0.00591555568906996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,128,0.0027022223091787765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,3072,0.015270221564504834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,4096,0.01904177831278907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,10240,0.018969777557584975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,2048,0.006625777731339137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,10240,0.03838133480813768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,10240,0.03925866550869412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,4096,0.05981422132915921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,10240,0.0069324444565508105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,3072,0.03927110963397556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,4096,0.46620268291897243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,12288,0.03157511022355821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,6144,0.019119999474949308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,4096,0.009352000223265754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,7168,0.013771555489963956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,5120,0.012278222375445895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,65536,0.40595467885335285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,3584,0.02781955401102702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,32,0.026392888691690233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,5120,0.008341333104504479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,32,0.010254222485754225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,128,0.0030906666070222855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,1536,0.042211555772357516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,128,0.002798222212327851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,128,0.004461333155632019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,1536,0.07980799674987793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,16384,0.09961422284444173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,16384,0.037257777320014104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,16384,0.03408888975779215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,5120,0.009279111193286048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,128,0.005074666606055366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,12288,1.6329999499850805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,65536,0.08323199881447686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,1536,0.006134222365087933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,1024,0.005115555392371284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,512,0.005588444570700328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,7168,0.013943110903104147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,2048,0.00700711127784517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,512,0.01901422275437249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,512,0.0057297779454125305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,7168,0.020936000678274367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,128,0.0027342221389214196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,3072,0.006302222195598815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,10240,0.09115822447670831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,2560,0.008298666940795051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,6144,0.025222222010294598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,64,0.0069520001610120135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,6144,0.30142490069071454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,1024,0.009731555150614845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,8192,0.02831022275818719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,8192,0.5672186745537652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,10240,0.028148445818159316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,3072,0.020030222005314298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,65536,0.13675199614630804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,1536,0.03601244423124526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,8192,0.01626755628320906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,768,0.007865777446164025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,12288,0.035440001222822405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,768,0.010339555641015371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,5120,0.011022222538789114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,6144,0.012050666742854647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,6144,0.01238044434123569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,256,0.005486222190989389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,3584,0.007874666816658443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,3584,0.018055111169815063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,128,0.006774222271309958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,7168,0.02180266711446974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,64,0.006520000182920032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,12288,0.008080888953473832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,3072,0.011056000159846412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,768,0.006686222222116258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,5120,0.01667022208372752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,512,0.007328000333574083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,256,0.0037760001917680106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,512,0.0037004442678557504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,128,0.0030746666921509635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,16384,0.08812089098824395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,768,0.004008000095685323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,10240,0.0237128883600235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,8192,0.028824888997607764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,1024,0.003954666770166821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,1536,0.00814666681819492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,768,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,1536,0.006799110935793982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,256,0.010270222193664974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,64,0.01682400041156345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,64,0.006649777707126405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,32,0.006441777778996362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,256,0.0032826666202810076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,65536,0.34967377450731063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,16384,0.6600542068481445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,512,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,2560,2.224069383409288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,65536,0.015805333852767944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,10240,0.343372450934516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,12288,0.01864088906182183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,1024,0.012077333198653327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,64,0.006177777631415262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,2560,0.006337777607970768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,768,0.0041804446114434134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,512,0.005809777726729711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,65536,0.17341421710120308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,6144,0.040822221173180476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,3072,0.013797333670987023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,2560,0.23082311948140463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,6144,0.013639999760521783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,768,0.005015111217896144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,10240,0.04900266726811727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,5120,0.06823199987411499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,64,0.0025439999169773525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,4096,0.028839111328125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,64,0.005156444592608346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,2048,0.007195555501514011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,3584,0.010694222317801582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,7168,0.012895999683274163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,4096,0.015111999379263984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,256,0.029567999972237483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,512,0.03367022342152066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,65536,0.22179999616410997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,5120,0.012093333734406365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,65536,0.8822133276197645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,5120,0.05867555406358507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,256,0.005647999958859549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,7168,0.01706044375896454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,1024,0.012089778151777057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,1536,0.008279999925030602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,256,0.006034666465388403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,10240,0.029009779294331867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,256,0.005196444276306364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,1536,0.024874667326609295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,768,0.0058453331391016645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,1024,0.006864000111818314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,16384,0.14135555426279703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,10240,0.03230133321550157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,2560,0.008367111285527548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,1024,0.009520000053776635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,10240,0.07779377698898315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,2048,0.005352889084153705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,1536,0.011959111524952782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,1536,0.005588444570700328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,768,0.09869422515233357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,2048,0.005296000176005893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,768,0.0046951112647851305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,3072,0.015896888242827523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,5120,0.0407626661989424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,768,0.0033662221911880705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,128,0.02844266759024726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,2560,0.010816000401973724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,1536,0.00943288869327969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,12288,0.020415999823146395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,8192,0.029793779055277508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,64,0.0024382221615976757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,256,0.0031155554784668815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,128,0.004507555729813046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,1536,0.006595555692911148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,768,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,3584,0.0722480018933614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,768,0.005110222018427319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,2048,0.006491555521885554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,4096,0.19822666380140516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,6144,0.023540443844265405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,1024,0.02923822071817186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,1024,0.11512445078955756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,1536,0.005701333284378052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,12288,0.6973750856187609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,1024,0.005364444520738389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,64,0.005192000004980299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,1024,0.010989333192507425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,10240,0.014534221755133735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,3072,0.01646399994691213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,1024,0.004208000169859992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,512,0.003882666842805015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,128,0.005195555587609609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,2048,0.025591111845440332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,128,0.0031111112071408164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,768,0.004856889032655292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,3584,0.05372977919048733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,3584,0.01741422216097514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,65536,0.11052799887127346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,3072,0.007037333316273159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,16384,0.006439111298984951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,32,0.004416000097990036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,2560,0.08609866433673435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,6144,0.01682222220632765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,3584,0.01574222246805827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,16384,0.03184533450338576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,16384,0.030606223477257624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,768,0.0051857779423395796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,4096,0.013409777647919126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,768,0.005468444277842839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,7168,0.041951109965642296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,512,0.003846222327815162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,5120,0.07912444406085543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,8192,0.024523556232452393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,2048,0.02602400051222907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,5120,0.012289777398109436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,2560,0.01389600005414751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,128,0.006991110742092133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,65536,0.0785342190000746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,512,0.003705777641799715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,4096,0.10995378096898396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,6144,0.014459555347760519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,128,0.009353777600659264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,128,0.002748444469438659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,128,0.002817777709828483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,128,0.046514666742748685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,512,0.010009778042634329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,32,0.0026702222724755607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,5120,0.0069066666894488865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,10240,0.022543110781245764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,7168,0.018256000346607633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,3584,0.01550222271018558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,8192,0.01183555523554484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,32,0.0033191111352708605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,12288,0.043113778034845986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,3072,0.018220444520314533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,256,0.010943999720944298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,10240,0.017486222916179232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,5120,0.024870221813519795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,3584,0.022256000174416438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,64,0.0026586666289303037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,1536,0.008909333083364699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,6144,0.03935199975967407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,7168,0.0454328887992435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,10240,0.09034844239552815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,768,0.010761777559916178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,16384,0.06355200211207072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,16384,0.018792889184421964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,1024,0.014929778046078153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,256,0.005416888743638992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,2048,0.006051555689838197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,6144,0.010249777800507015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,2560,0.008372444245550368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,10240,0.03846577803293864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,3072,0.07297244336869982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,2560,0.0070977773931291364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,8192,1.3812132941351996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,1024,0.006728000111050076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,16384,0.04289688997798496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,2560,0.00777511133088006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,6144,0.012744888663291931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,8192,0.02375999920898014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,2048,0.01564711166752709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,768,0.0053839998112784485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,5120,0.007043555378913879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,65536,0.09715911414888169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,768,0.005710222240951326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,256,0.0029537777105967202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,16384,0.02143200072977278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,12288,0.05250933435228136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,10240,0.051093333297305636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,1024,0.10043288601769342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,128,0.007206222249401941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,256,0.0036106668412685394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,2560,0.007288888924651676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,7168,0.031239999665154353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,2560,0.016897777716318767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,768,0.005673777725961473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,1536,0.010785777535703448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,768,0.013204444613721637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,512,0.0053395554423332214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,8192,0.03207555413246155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,16384,0.03792355457941691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,3584,0.5268293486701118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,16384,0.4009839958614773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,1024,0.006640888750553131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,65536,9.883949279785156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,5120,1.0253243976169162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,6144,0.014567111929257711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,768,0.014734221829308404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,1024,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,2048,0.008907555705971187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,65536,0.4103822178310818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,16384,0.024043555061022442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,768,0.00554577757914861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,6144,0.5866897900899252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,256,0.005522666705979242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,64,0.003489777859714296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,16384,0.03963022099600898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,1024,0.005047111047638787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,12288,0.015072888798183866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,64,0.05514311128192478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,5120,0.009155555731720394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,12288,0.4834159745110406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,5120,0.009226666556464301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,8192,0.008428444465001425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,10240,0.024422221713595923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,7168,0.018479110466109384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,10240,0.013182222015327878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,7168,0.012664889295895895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,1536,0.00887022250228458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,768,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,2048,0.0057839999596277876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,10240,0.0387644436624315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,10240,0.05591377947065565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,65536,1.7868266635470922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,8192,0.021890666749742296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,5120,0.010079111489984725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,3584,0.03990222348107232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,65536,1.1063928604125977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,2560,0.016335111525323655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,64,0.005647999958859549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,128,0.006073777874310811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,7168,0.09134222401512994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,10240,0.025063999825053748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,256,0.008357333640257517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,512,0.0032666667054096856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,7168,0.026615111364258662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,6144,0.011124444504578909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,3072,0.021631111701329548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,7168,1.6358932918972438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,1024,0.00925511121749878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,3584,0.031665778822369046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,3072,0.01551555593808492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,3584,0.007263999846246507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,10240,0.05529600050714281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,65536,0.2528088887532552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,2048,0.011130666567219628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,512,0.008784888519181145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,7168,0.03172711200184292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,128,0.004135999828577042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,1024,0.012131555212868584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,4096,0.017961778574519686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,3584,0.010760889285140567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,12288,0.016551999582184684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,5120,0.01309866706530253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,8192,0.03438755538728502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,4096,0.013423110875818463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,768,0.005072000126043956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,64,0.0026444445053736367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,768,0.004033777862787247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,65536,0.428056001663208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,1536,0.005440000030729506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,3072,0.006888888776302338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,16384,0.02327288852797614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,2560,0.011063111325105032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,65536,0.06746133168538411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,1536,0.14432266023423937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,3584,0.007720000214046902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,5120,0.01851111153761546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,6144,0.021831111775504217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,1536,0.005688888745175467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,7168,0.021733333667119343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,64,0.008287111090289222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,2048,0.010260444548394945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,128,0.005482666608360078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,256,0.024268445041444566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,4096,0.01903733279969957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,1536,0.0247662216424942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,768,0.013883555928866068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,128,0.006041777630647023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,8192,0.044983112149768405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,65536,0.31424087948269314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,512,0.004651555584536659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,3584,0.008615111311276754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,3584,0.015198222464985318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,4096,0.010449777874681683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,16384,0.037251555257373385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,256,0.00638133328821924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,65536,0.07250489128960504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,128,0.007107555866241455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,16384,0.03915822174814012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,4096,0.1784977780448066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,128,0.004017777740955353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,512,0.004886222382386525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,12288,0.035904000202814736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,768,0.005894222193294101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,256,0.007668444679843054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,512,0.0032417778339650896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,3584,0.009471110999584198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,32,0.003522666792074839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,32,0.002838222309947014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,16384,0.29053065511915416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,6144,0.19440800613827178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,512,0.0030871110243929755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,10240,0.06798311074574788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,3584,0.02718222141265869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,2560,0.010799999866220685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,3584,0.018062222335073683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,4096,0.02455555564827389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,2560,0.007184888753626082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,512,0.005504000104135937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,7168,0.0325617790222168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,2048,0.023215999205907185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,7168,0.030485332012176514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,64,0.0026835555003749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,512,0.004470222112205294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,16384,0.7008390956454806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,4096,0.019464888506465487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,1536,0.006303999986913469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,1024,0.006726222319735422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,2048,0.02938666608598497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,32,0.010679110884666443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,1536,0.0068693334857622785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,3072,0.07016177972157796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,7168,0.09893600145975749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,256,0.009136888715955947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,12288,0.017456000049908955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,64,0.029827555020650227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,1024,0.07681955231560601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,32,0.06049955553478665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,6144,0.018953777021831937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,1024,0.005895111295912001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,6144,0.08670488993326823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,12288,0.40913067923651797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,1536,0.009815110928482479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,8192,0.007479999628331926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,65536,0.06988266441557142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,3584,0.01554844445652432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,128,0.12197244167327881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,16384,0.027582221561008032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,64,0.003162666741344664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,768,0.011100444528791638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,16384,0.17240799797905815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,3072,0.02014044423898061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,6144,0.011125333607196808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,6144,0.0120666664507654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,1536,0.011735999749766456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,8192,0.06605422496795654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,768,0.024488000406159297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,7168,0.02279377811484867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,768,0.005557333429654439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,10240,0.017421333326233756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,512,0.007139555282062954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,7168,0.006261333409282897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,10240,0.007105777661005656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,128,0.006074666563007567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,1536,0.0162462227874332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,65536,0.20411022504170737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,8192,0.089336891969045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,12288,0.006524444454246097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,65536,0.1311475568347507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,16384,0.02130222154988183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,768,0.0045422220395671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,768,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,3072,0.016352888610627916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,16384,0.054662221007876925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,5120,0.0125608891248703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,1024,0.009381333159075843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,16384,0.02516977820131514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,3072,0.016511110795868766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,5120,0.011755555868148804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,2048,0.02546311087078518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,12288,0.9031591415405273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,65536,0.11790577570597331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,128,0.00407911092042923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,2560,0.4878942171732585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,128,0.008108444511890411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,4096,0.009857777920034196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,12288,0.055847108364105225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,65536,1.5776835547553167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,768,0.0074026667409473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,3072,0.02420266634888119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,128,0.032280001375410296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,4096,0.009015999734401703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,12288,0.018138666947682697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,768,0.004368888835112254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,12288,0.02681866619322035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,4096,0.03312355611059401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,16384,0.05446222093370226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,5120,0.00888355573018392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,4096,0.03349511159790887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,4096,0.02981688910060459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,3072,0.13675910896725124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,512,0.004618666652176115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,12288,0.04744355546103584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,3584,0.010820444259378644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,1536,0.007254222200976477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,256,0.003162666741344664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,256,0.006506666541099548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,5120,0.03718488746219211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,3584,0.11955911583370632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,3072,0.015028445257080926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,768,0.004383111165629493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,256,0.004523555437723796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,3072,0.007159111400445302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,512,0.007855110698276097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,256,0.007178666690985362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,3584,0.02087644404835171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,128,0.005161777966552311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,8192,0.016764443781640794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,4096,0.019285332825448778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,32,0.002899555489420891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,65536,0.07315911187065972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,128,0.006136000156402588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,10240,0.014334221680959066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,2048,0.018545778261290658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,12288,0.13023555278778076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,1536,0.004305777864323722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,16384,0.02808088726467556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,256,0.00565244464410676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,2560,0.007175111108356052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,3072,0.013260444833172692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,8192,0.019168888529141743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,512,0.003439999909864532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,2048,0.011874666644467248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,512,0.0038693332009845306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,32,0.008755555583371056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,5120,0.01091733326514562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,12288,0.04804177747832405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,64,0.0032319999817344877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,768,0.004795555439260271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,64,0.005315555466545953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,8192,0.04571733209821913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,16384,0.05305066704750061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,65536,0.08879644340938991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,10240,2.3670461442735458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,6144,0.01071111112833023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,2560,0.1083128915892707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,10240,0.5977653397454156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,32,0.007879111501905654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,65536,0.7274933391147189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,2048,0.006050666587220297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,2048,0.009674666656388177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,2560,0.006155555446942647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,256,0.0047475554876857335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,1536,0.007419555551475949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,256,0.06587644418080647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,5120,0.027561777167850073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,7168,0.012936000194814471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,10240,0.02647200061215295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,12288,1.1929386986626518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,768,0.004795555439260271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,1536,0.004385777645640903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,10240,0.2531031237708198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,2048,0.011566222541862063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,12288,0.040864000717798867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,64,0.004295999805132548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,6144,0.11402400334676106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,10240,0.016515556308958266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,4096,0.05668355358971489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,10240,0.09524977869457668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,7168,0.011240888800885944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,128,0.0116462219092581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,12288,0.4517857763502333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,1024,0.009966221948464712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,10240,0.041505777173572116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,10240,1.1224293178982205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,7168,0.013236444857385425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,2560,0.011688888900809817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,6144,0.014263111684057446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,256,0.006060444646411472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,1536,0.009020444419648912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,4096,0.008454221818182204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,16384,0.008370666868156856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,256,0.00702133360836241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,1024,0.00888888869020674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,256,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,768,0.012439111040698158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,10240,0.032223112053341336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,4096,0.10130400127834743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,8192,0.01585422290696038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,3584,0.010201777848932479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,64,0.01977688901954227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,6144,0.01831822262869941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,10240,0.017811555001470778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,5120,0.01129955550034841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,1536,0.00480266660451889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,3072,0.011035555766688453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,6144,0.013648000028398303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,64,0.0053004444473319584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,512,0.005165333135260476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,2560,0.00831288927131229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,4096,0.1636755598915948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,7168,0.01423466702302297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,1024,0.005963555640644497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,7168,0.19147200054592559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,512,0.005513777749405966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,4096,0.01127822200457255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,12288,0.4984560012817383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,2048,0.02608000073168013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,512,0.008347555167145198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,768,0.008285333712895712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,64,0.005872888697518243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,65536,0.07390666670269437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,32,0.005852444304360284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,6144,0.011908444265524546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,4096,0.032800889677471586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,7168,0.020059555768966675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,3584,0.01181777815024058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,8192,0.015976889265908133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,128,0.004323555363549126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,8192,0.1395475599500868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,10240,0.05762222078111437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,2048,0.008945778012275696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,6144,0.01399466726515028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,2048,0.007469333708286285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,8192,0.011880888707107969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,3072,0.0074728892909155945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,7168,0.04237066706021627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,32,0.008933333059151968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,3072,0.016524444023768108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,128,0.004417777889304691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,128,0.004294222427739037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,65536,13.208112080891928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,1536,0.06016888883378771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,12288,0.010588444769382477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,2048,0.012839111189047495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,256,0.00664533343580034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,10240,0.008564444051848518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,256,0.0047626665068997275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,5120,0.10787289010153876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,16384,0.05893333090676201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,4096,0.015669332610236276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,256,0.004713777866628435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,5120,0.007551111280918121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,4096,0.036574221319622464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,512,0.005648889061477449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,512,0.005125333451562458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,1536,0.03247733248604669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,65536,0.3962257703145345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,3584,0.017134222719404433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,6144,0.013170666992664337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,2048,0.005271111097600725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,5120,0.008346666892369589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,10240,0.01426488823360867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,8192,0.011085333095656501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,65536,0.06690755817625257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,5120,0.05774844355053372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,4096,0.025236444340811834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,5120,0.00977244476477305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,32,0.005403555515739653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,3584,0.008242666721343994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,768,0.026639110512203638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,10240,0.059112886587778725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,65536,1.2387350930107963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,12288,0.2596204545762804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,7168,0.008984000318580205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,128,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,256,0.00407911092042923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,2048,0.007135999699433644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,65536,0.24968356556362578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,8192,0.09292444255616929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,64,0.02349066734313965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,12288,0.3028755452897814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,128,0.041264888313081526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,1024,0.006220444622966979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,64,0.00618488879667388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,12288,0.018037334084510803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,5120,0.05792178048027886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,5120,0.013847110999955071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,7168,0.015018666783968607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,256,0.0058168888919883305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,512,0.011701333026091257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,8192,0.028839111328125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,3072,0.007678222325113084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,768,0.009647111097971598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,768,0.005703111075692707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,8192,0.028520888752407495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,2560,0.006344888773229387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,1536,0.03199288911289639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,768,0.007805333369308048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,1024,0.004008000095685323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,12288,0.011209777659840055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,768,0.004046222401989831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,3072,0.009926222264766693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,1536,0.011721777419249216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,2560,0.03811288873354594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,1024,0.009735999835862054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,64,0.002946666752298673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,5120,0.02481511069668664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,64,0.05620533227920532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,3072,0.019256888164414298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,128,0.004809777769777509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,4096,0.010873777998818291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,16384,0.05526133378346761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,64,0.008247999681366814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,8192,0.013208000196350945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,2560,0.11617778407202826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,512,0.004435555388530095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,8192,0.2627404530843099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,8192,0.03990222348107232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,8192,0.02948088778389825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,128,0.0033457777980301115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,2560,0.013600889179441663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,32,0.004974222017659081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,2048,0.010211555494202508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,3584,0.010121777653694153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,8192,0.043415112627877124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,2048,0.012999999854299756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,10240,0.014398222168286642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,768,0.00507377791735861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,12288,0.12608888414171007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,10240,0.05273688832918803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,128,0.005202666752868229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,256,0.00546577779783143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,1024,0.007601777712504069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,768,0.008184889124499427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,768,0.015063999427689446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,1536,0.011584888729784222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,2560,0.014178666803571912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,3584,0.008616888688670265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,768,0.003825777934657203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,16384,0.10803111394246419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,7168,0.05846933523813883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,3072,0.008742222355471717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,6144,0.029924445682101782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,7168,0.6779893239339193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,8192,0.007461333440409766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,2048,0.03089155422316657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,16384,0.12761510743035212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,1536,0.004831111265553368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,768,0.006180444525347815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,7168,0.025669333007600572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,12288,0.016884444488419425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,16384,0.0465866658422682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,12288,0.033519999848471746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,1024,0.01223111069864697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,65536,0.06565422481960721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,768,0.0052577778697013855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,16384,0.294739564259847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,4096,0.13843111197153726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,768,0.003941333128346337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,128,0.00544799988468488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,128,0.0029786665820413162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,2560,0.008932444784376357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,10240,0.042034665743509926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,1024,0.005984000033802456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,12288,0.043825778696272105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,1536,0.007693332930405934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,768,0.006239111224810283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,7168,0.02051555613676707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,65536,0.061786664856804736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,3584,0.01869422197341919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,3584,0.018061333232455783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,64,0.0033395555284288195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,5120,0.04574933316972521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,512,0.00721866637468338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,1024,0.00600355573826366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,768,0.008008889026112026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,12288,0.16891911294725206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,3072,0.6345155504014757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,16384,1.5422222349378798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,128,0.025083555115593806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,3584,0.01296533313062456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,2048,0.007575111256705389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,10240,0.019739554988013375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,2048,0.005487111293607288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,768,0.006067555397748947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,3584,0.31824533144632977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,2048,0.009470221896966299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,12288,0.04784977767202589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,3584,0.016034666034910414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,512,0.004328888737493091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,3584,0.029208001163270738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,512,0.005215999980767568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,6144,0.006083555519580841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,10240,0.08470311429765488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,4096,0.007606222397751278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,10240,0.015606222881211175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,7168,0.02988622254795498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,3584,0.01256533298227522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,16384,0.022330666581789654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,64,0.00254844439526399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,16384,0.0528400010532803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,2048,0.005702222386995952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,5120,0.008884444832801819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,10240,0.013140444126394061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,768,0.005169777820507686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,64,0.006439111298984951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,4096,0.008087111016114553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,2048,0.007145778172545963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,4096,0.013558222187889947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,32,0.004968889057636261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,1024,0.00598933340774642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,2560,0.01420177850458357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,128,0.005256888767083486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,1024,0.011920000116030375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,7168,0.09299911393059625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,6144,0.02683111031850179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,3584,0.014920888675583733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,1024,0.03411644366052415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,1024,0.006180444525347815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,3072,0.010173333187898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,2560,0.022112000319692824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,512,0.005254222287072076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,6144,0.006298666612969504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,768,0.006110222389300664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,2048,0.01293333371480306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,10240,0.5002853075663248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,3584,0.01145600030819575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,1024,0.057095997863345675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,256,0.007377777662542131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,2048,0.016536888149049547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,512,0.011752000285519494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,8192,0.03121333320935567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,2048,0.0188426673412323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,8192,0.03324710991647508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,1536,0.04853689008288913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,6144,0.012582222620646158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,2560,0.01927199959754944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,8192,0.03064799971050686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,7168,0.020544000797801547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,16384,0.02084000077512529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,65536,0.07426222165425618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,5120,0.021189333664046392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,2560,0.0467564430501726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,7168,0.00963644435008367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,1024,0.006578666468461354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,5120,0.015840889679061044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,6144,0.06180266539255778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,2048,0.1022533310784234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,7168,0.021333333518770006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,128,0.003946666502290302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,10240,0.013432000246312885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,65536,0.41597512033250594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,1024,0.006034666465388403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,10240,0.020802666743596394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,2560,0.008307555483447181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,2048,0.008088000118732452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,10240,0.027238221632109746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,12288,0.011527111132939657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,256,0.004791111167934206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,7168,0.013354666531085968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,32,0.005428444594144821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,6144,0.011241777903503843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,64,0.04516355527771843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,512,0.004880889008442561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,65536,0.1917217837439643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,3072,0.01519555515713162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,3072,0.006899555524190267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,2048,0.02480799953142802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,3072,0.007791111038790808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,2560,0.007079111205206976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,512,0.004937777916590373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,3584,0.0069431112044387394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,4096,0.15914844142066106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,768,0.03906133439805772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,6144,0.012667555775907306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,3072,0.04455555478731791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,10240,0.04413066638840569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,10240,0.03765777746836344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,10240,0.011036444041464063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,3072,0.01646311084429423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,128,0.016556445095274184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,256,0.0070497774415545994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,2048,0.014546665880415173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,1024,0.03336888882848952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,256,0.005409777578380372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,2048,0.009179555707507664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,6144,0.009564444422721863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,7168,0.014928888943460254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,128,0.003952888978852166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,7168,0.9249760309855143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,7168,0.019765333996878732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,8192,0.008577778107590145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,1024,0.01640977793269687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,768,0.018722666634453666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,12288,0.01994400057527754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,768,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,3072,0.012224888636006249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,7168,0.0590897798538208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,512,0.012892444100644855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,256,0.006673777682913675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,1536,0.004903111192915174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,2048,0.00871288859181934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,65536,0.021538666552967493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,4096,0.007658667034573025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,1024,0.01072088877360026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,768,0.007927999728255803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,2048,0.01591999994383918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,5120,0.009337777892748514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,10240,0.0625964469379849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,512,0.0162782222032547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,6144,0.006505777852402792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,16384,0.018730666902330186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,256,0.0030533334033356774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,10240,0.03430577781465318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,10240,0.44302932421366376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,8192,0.04665333363744947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,128,0.0058551111982928375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,12288,0.05070844292640686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,1024,0.028189334604475234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,512,0.005445333404673471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,3072,0.41291554768880206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,512,0.00627377794848548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,8192,0.013648889131016202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,256,0.00582755563987626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,16384,0.06137688954671224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,1536,1.2156844668918185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,16384,0.2971617645687527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,8192,0.02496533261405097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,256,0.004247111164861255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,10240,0.04513066675927904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,5120,0.011600889265537262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,32,0.003136000078585413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,2560,0.0123831108212471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,512,0.013185777597957186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,6144,0.01573422220018175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,512,0.0032835555159383347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,65536,0.09470311138365005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,32,0.010335110955768162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,16384,0.06407466861936781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,6144,0.13867289490169948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,256,0.005500444521506627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,16384,0.02609955602222019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,2560,0.015844444433848064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,5120,0.006598222172922558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,12288,0.032249778509140015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,2560,0.008075555165608725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,2560,0.00796977761718962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,1024,0.009392889009581672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,10240,0.02122488949033949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,2560,0.005501333210203383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,512,0.006242666807439592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,6144,0.017151999804708693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,32,0.005386666705211003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,5120,0.024274667104085285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,12288,0.017684444785118103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,10240,0.03264888789918687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,64,0.0047644442982143825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,768,0.006029333505365584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,4096,0.017288888494173687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,1536,0.005320888840489917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,10240,0.059600002235836454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,5120,0.09198755688137478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,32,0.002609777781698439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,5120,0.08908977773454453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,16384,0.8744782341851128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,3072,0.020601777566803824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,12288,0.019173334042231243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,6144,0.03584444522857666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,64,0.03192977772818671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,1536,0.008959111240175035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,3072,0.012686221963829465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,512,0.004566222015354368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,12288,0.03009777930047777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,65536,0.38509511947631836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,2048,0.01718577742576599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,3584,0.0291884442170461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,16384,0.09149155351850723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,65536,0.0700755582915412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,5120,0.012094222009181976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,7168,0.009211555951171452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,128,0.004511111312442356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,3072,0.009167110754383935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,2048,0.00979466653532452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,2560,0.027686221732033625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,8192,0.012984889249006907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,3072,0.013167111410035027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,16384,0.04784889022509257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,12288,0.29231378767225474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,4096,0.13208622402615017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,16384,1.1120177374945748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,3584,0.012503999802801343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,1536,0.013179555535316467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,7168,0.018018666240904067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,4096,0.010969777901967367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,1536,0.009565333525339762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,16384,0.27638400925530326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,3584,0.07867555485831366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,3584,0.007296889192528195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,5120,0.02146311104297638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,4096,0.013933334085676404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,7168,0.009111111362775167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,4096,0.1794666714138455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,12288,0.051098667913013034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,3584,0.08269866969850329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,3072,0.012059555285506778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,2560,0.01311555587583118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,7168,0.020956445071432326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,4096,0.009340444372759925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,6144,0.02804533309406704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,5120,0.010840000377760993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,512,0.008714666797055138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,768,0.0778240031666226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,7168,0.01442488862408532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,65536,0.2491226726108127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,5120,0.09336533149083455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,128,0.005200888961553574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,512,0.008749333520730337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,12288,0.3940897782643636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,5120,0.052226665947172374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,6144,0.010861333045694562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,1024,0.006940444310506185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,1536,0.007010666860474481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,65536,0.06355555852254231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,128,0.06087377998563978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,16384,0.02145066691769494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,65536,2.8037804497612844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,128,0.0033404444240861465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,5120,0.04518844352828132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,2560,0.006590222318967183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,3072,0.03523022267553542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,3584,0.017567111386193168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,1536,0.021405332618289526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,12288,0.8928933673434787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,3584,0.015584889385435315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,1024,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,8192,0.013556444810496436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,1024,0.009847111172146266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,3584,0.009001777403884465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,2560,0.011144888897736868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,256,0.005579555614127054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,10240,0.013379555609491138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,4096,0.13601956102583143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,10240,0.1946551137500339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,256,0.005183111048407025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,5120,0.011056000159846412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,5120,0.01725955473052131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,4096,0.007195555501514011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,128,0.002817777709828483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,16384,0.04475466741455925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,256,0.0075804442167282104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,7168,0.02755555510520935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,32,0.012159999873903064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,768,0.00889955543809467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,64,0.002777777819169892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,256,0.004722666823201709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,7168,0.029314666986465454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,1024,0.004680888934267892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,8192,0.27224710252549916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,1024,0.0063733334342638654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,3584,0.012464889221721225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,1024,0.004850666556093427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,1536,0.0047591109242704184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,7168,0.03148266673088074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,4096,0.023056889573733013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,1536,0.007948444121413762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,1536,0.014012444350454541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,10240,0.03360088997417026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,16384,0.06495377752516004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,6144,0.028712888558705647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,256,0.005380444642570283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,512,0.0188044442070855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,4096,0.011585777832402123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,6144,0.7808773252699109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,3584,0.009974222216341231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,16384,0.049887110789616905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,7168,0.013621333572599622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,4096,0.0069102222720781965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,10240,0.016338666280110676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,7168,0.011543110840850405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,64,0.011504888534545898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,512,0.021928000781271193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,256,0.008393777741326226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,256,0.004229333251714706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,1536,0.006052444378534953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,1536,0.006009777800904379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,8192,0.052935110198126904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,6144,0.43868088722229004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,2048,0.006686222222116258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,10240,0.03479555580351088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,3072,0.01676977839734819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,65536,0.22164889176686606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,512,0.00583022211988767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,5120,0.37790311707390684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,6144,0.021749332547187805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,768,0.005889777921968036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,3584,0.049950222174326576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,64,0.00300177786913183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,64,0.008928000099129146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,768,0.00980977796845966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,2560,0.007568000091446771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,256,0.004406222038798862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,32,0.005466666486528184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,2048,0.006968888971540663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,256,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,7168,0.013412444127930535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,4096,0.010841777755154504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,5120,0.2023031049304538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,6144,0.015104000767072042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,8192,0.06924977567460802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,7168,0.01223288890388277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,12288,0.03199733297030131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,4096,0.01976711054642995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,1536,0.006395555618736479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,7168,0.006085333310895496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,8192,0.025172443853484258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,3072,0.014208889669842191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,64,0.005165333135260476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,1536,0.005371555685997009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,1024,0.013306666579511432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,768,0.00619555554456181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,512,0.007466666400432587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,10240,0.02027555637889438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,32,0.0031946665710873077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,1024,0.01146933353609509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,4096,0.01605333387851715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,16384,0.025819554924964905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,8192,0.019088000059127808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,3584,0.008072888685597314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,16384,0.02277155551645491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,4096,0.01774577796459198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,5120,0.023402666052182514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,3072,0.008390222158696916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,8192,0.06550577614042494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,4096,0.03781066669358148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,3072,0.015282667345470853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,3072,0.0885537796550327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,5120,0.05905866622924805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,6144,0.08449244499206543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,2560,0.047800001170900136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,2048,0.00943999985853831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,65536,0.09936800267961289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,128,0.005025777965784073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,65536,0.12210755878024632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,512,0.1565706729888916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,1536,0.007998222278224098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,32,0.006953777538405524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,10240,0.018752889500723947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,16384,0.0229768885506524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,3072,0.007708444363541073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,10240,0.01720355616675483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,65536,0.16458843814002141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,65536,8.18955315483941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,3584,0.043503112263149686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,256,0.00535733335547977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,64,0.0026168889469570587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,3584,0.013295999831623502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,32,0.0027502222607533136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,512,0.006504000061088138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,10240,0.026957333087921143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,2048,0.007711999946170383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,6144,0.016896888613700867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,4096,0.11454044448004828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,8192,0.04615733358595106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,4096,0.020404444800482858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,12288,0.08894399801890056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,64,0.004463110946946674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,5120,0.016487111647923786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,4096,0.01198755535814497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,768,0.008576889004972246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,7168,0.22511111365424263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,4096,0.008963555925422246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,65536,0.14021510548061794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,768,0.009310222334331935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,65536,0.066666669315762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,256,0.004751111070315043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,8192,0.04271644353866577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,2048,0.02019733356104957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,7168,0.050695998801125415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,1024,0.005067555440796746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,10240,0.31341245439317494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,1536,0.004574222283230888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,16384,0.04089422358406915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,2560,0.008609777523411645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,6144,0.015561777684423657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,256,0.020822222034136455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,3584,0.10148444440629746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,7168,0.016806221670574613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,256,0.01900622248649597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,5120,0.1583013269636366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,65536,0.08702222506205241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,8192,0.0120319997270902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,768,0.00499288903342353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,1536,0.01036888857682546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,768,0.0239955551094479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,2048,0.015671999918089975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,16384,7.055952707926433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,10240,0.04572977622350057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,4096,0.01421955558988783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,16384,0.026940445105234783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,2560,0.008632000121805403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,768,0.007423111134105259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,3072,0.012774222426944308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,64,0.006171555568774541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,512,0.05429244372579786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,10240,0.01794311073091295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,5120,0.011190222369299995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,65536,0.2756853368547228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,256,0.006075555665625467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,7168,0.19257067309485543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,12288,0.050238221883773804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,8192,0.27748266855875653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,16384,0.02323644525474972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,64,0.008071111308203803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,3072,0.009069333473841349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,4096,0.013120888835854001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,6144,0.012122666670216454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,10240,0.029080000188615587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,128,0.005541333307822545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,4096,0.01054933336046007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,4096,0.018791110979186166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,3584,0.006439111298984951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,6144,0.014304000470373364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,3584,0.17541778087615967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,12288,0.023535999986860488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,16384,0.08838044272528754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,5120,0.5702186690436469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,3584,0.007912889122962952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,64,0.002577777744995223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,4096,0.13063377804226345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,256,0.004474666797452503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,2560,0.05039199855592516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,2560,0.02814488940768772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,8192,0.01980888843536377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,3584,0.18124711513519287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,1024,0.004200889004601372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,5120,0.008968888885445064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,65536,0.06446044312583075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,1536,0.0063288890653186375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,65536,0.043487111727396645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,512,0.01632355484697554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,128,0.004137777619891697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,512,0.009082666701740688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,512,0.0034391110142072043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,2048,0.007096889118353526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,7168,0.023896889554129705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,512,0.006769777586062749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,65536,1.1477111180623372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,2560,0.006858666737874349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,2048,0.022062222162882488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,1536,0.012123555772834353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,12288,0.021469333105617102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,3584,0.07135288582907783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,2048,0.01386755539311303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,8192,0.007041778001520369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,8192,0.02294844388961792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,3584,0.009533333281675974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,4096,0.013872000078360239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,3072,0.09742133484946357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,65536,0.543175114525689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,16384,0.748927116394043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,7168,0.012359999948077731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,3584,0.014165333575672574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,2560,0.0444453325536516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,3072,0.006814222368929122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,2048,0.017916444275114272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,256,0.0032542221662071017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,768,0.012637333737479316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,6144,0.018908444378111098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,1024,0.05945777893066406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,12288,1.8632916344536676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,7168,0.017417778571446735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,12288,0.023800000548362732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,65536,0.2572408782111274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,8192,0.05661155780156454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,8192,0.1736826631757948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,256,0.0074035558435651995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,5120,0.013989332649442883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,2048,0.004931555440028508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,512,0.005092444519201915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,12288,0.1532764434814453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,128,0.0027217778066794076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,8192,0.027650665905740526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,6144,0.014500444134076437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,1024,0.004422222160630756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,12288,0.07407466570536296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,6144,0.02404888967672984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,1024,0.007715555528799693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,6144,0.3066177897983127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,512,0.0037928890022966596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,256,0.008937777744399177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,10240,0.03767022159364488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,3072,0.0092622223827574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,512,0.006016888966162999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,65536,0.07247377766503228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,16384,0.044497778018315635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,768,0.0047582222355736625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,8192,0.1429333289464315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,64,0.007314666277832455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,6144,0.01295644458797243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,12288,0.01871555546919505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,4096,0.014818666709793938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,6144,0.008280889027648503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,8192,0.016700444949997794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,10240,0.016629333297411602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,4096,0.040622221099005804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,2560,0.021174222230911255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,65536,0.2217573324839274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,3072,0.20417511463165283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,64,0.002895111011134254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,10240,0.5625893274943033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,3584,0.013379555609491138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,4096,0.023050665855407715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,7168,0.010994666980372535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,1536,0.006753777878151999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,8192,0.36908711327446836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,7168,0.04180977741877238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,3584,0.008660444782839881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,256,0.005912888795137405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,128,0.0032666667054096856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,1536,0.014012444350454541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,16384,0.10131555795669556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,256,0.004571555389298333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,10240,0.017383111847771537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,256,0.005335999859703912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,768,0.01961688862906562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,3072,0.009273777405420939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,2048,0.011708444191349877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,6144,0.02032888929049174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,512,0.005475555443101459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,10240,0.09047644668155247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,2048,0.00960888879166709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,3072,0.007241778075695038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,6144,0.01529600057337019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,65536,0.014448000325096978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,5120,0.12439110543992783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,8192,0.022454222043355305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,8192,0.03835733400450812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,1024,0.011735111474990845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,6144,0.014730667074521383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,3072,0.01126933346192042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,7168,0.032179554303487144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,256,0.01572800013754103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,3072,0.008028444316652086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,2048,0.00981155534585317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,2560,0.006085333310895496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,3072,0.013117333253224691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,4096,0.19278932942284477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,6144,0.20474489529927573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,6144,0.04076000054677328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,2560,0.008169777691364288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,3072,0.02791733211941189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,256,0.007070221834712558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,6144,0.016899555921554565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,4096,0.009156444834338294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,10240,0.017479999197853934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,8192,0.17092622650994194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,2560,0.047341333495246045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,1536,0.006708444406588872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,3072,0.042066666815016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,3072,0.10935022433598836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,65536,0.10485333204269409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,65536,0.22375999556647408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,8192,0.016537777251667447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,512,0.0043333334227403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,256,0.005410666680998272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,32,0.0053546664615472155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,4096,0.011235555840863122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,8192,0.19955733087327746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,3072,0.010696888797812991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,7168,0.016208888755904306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,10240,0.036694222026401095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,256,0.003302222324742211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,1024,0.008229333493444655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,768,0.031622221072514854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,32,0.006188444379303191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,16384,0.14248889022403294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,256,0.009153777526484596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,1024,0.007779555188284979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,8192,0.15730044576856825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,256,0.015691555208630033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,512,0.0035786665976047516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,7168,0.13852800263298884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,128,0.004091555459631814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,65536,0.06896177927652995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,2560,0.05835377507739597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,7168,0.016950221525298226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,8192,0.045801778634389244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,32,0.0027546667390399506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,7168,0.02700977855258518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,768,0.18432177437676323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,12288,0.017791999710930716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,32,0.0027297778675953546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,32,0.002587555597225825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,128,0.003117333269781537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,1536,0.004394666602214177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,12288,0.10195555289586385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,8192,0.07854933208889432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,12288,0.039664887719684176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,3072,0.02730133301681942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,6144,0.025419554776615564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,256,0.0033662221911880705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,12288,0.1078382200664944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,2560,0.08089955647786458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,6144,0.006280888699822956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,768,0.013533333109484779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,7168,0.026151999831199646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,2560,0.005983110931184556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,128,0.0032844444115956626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,768,0.017264000243610807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,4096,0.6628906461927626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,8192,0.04236444499757555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,12288,0.03822133276197646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,768,0.007151111132568783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,64,0.04292266567548116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,64,0.0035911111368073356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,16384,0.02701866626739502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,3072,0.026416889495319788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,128,0.006308444258239534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,6144,0.03547999925083584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,8192,0.07552978065278795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,512,0.005380444642570283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,16384,0.15125332938300237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,7168,0.06917688581678602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,6144,0.032383110788133405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,768,0.003936000168323517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,5120,0.10126399993896484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,32,0.005026666654480828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,3072,0.01144088887506061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,32,0.007793777518802219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,2560,0.08127822478612264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,65536,0.16876622041066489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,3584,0.13934489091237387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,2560,0.029513776302337646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,2560,0.030652443567911785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,2560,0.020827554994159274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,6144,0.20027732849121094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,2048,0.012653333445390066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,8192,0.06937777996063232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,768,0.0038844446341196695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,32,0.005219555563396878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,2048,0.004955555415815777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,256,0.0033457777980301115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,64,0.004984000076850255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,768,0.0063680000603199005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,10240,0.015959110524919298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,5120,0.03536266750759549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,65536,0.06956799825032552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,10240,0.02641244398223029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,768,0.02411288850837284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,32,0.002578666640652551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,12288,0.018594667315483093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,768,0.18954311476813424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,2048,0.020090666082170274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,8192,0.013045333325862885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,1024,0.02347822156217363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,5120,0.014078223043017916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,7168,0.0451546675629086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,2048,0.0052328887912962176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,10240,0.013467555244763693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,2560,0.0087013335691558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,32,0.004559111264016894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,12288,0.017958222164048087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,4096,0.028813332319259644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,7168,0.013623110949993134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,3584,0.006834666762087081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,768,0.003630222131808599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,768,0.023909333679411147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,10240,0.022983110613293115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,65536,0.04042399923006693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,768,0.008864888714419471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,10240,0.01311822235584259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,8192,0.05837066968282064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,8192,0.01311644415060679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,5120,0.024683554967244465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,12288,0.015087111128701104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,5120,0.031684445010291204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,768,0.005581333405441708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,10240,0.5289688640170628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,2560,0.01572711103492313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,6144,0.01918044520748986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,256,0.009546666509575313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,5120,0.01072622256146537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,12288,0.018428444862365723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,12288,0.042875554826524526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,1536,0.005680000202523337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,16384,0.050400889582104154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,16384,0.009167999857001835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,8192,0.009085333181752099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,1024,0.00853066643079122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,256,0.003375111147761345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,64,0.0027502222607533136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,65536,0.08182933595445421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,6144,0.0273635552989112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,8192,0.01109333336353302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,12288,0.006503110958470239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,16384,0.15620711114671496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,2048,0.005789333333571752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,10240,0.025597333908081055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,1024,0.006019555446174409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,64,0.002712888850106133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,128,0.005889777921968036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,768,0.0047315553658538395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,4096,0.11617955896589492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,128,0.0033697777738173804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,2560,0.04348177711168925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,64,0.004878222114510006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,7168,0.013716444373130798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,1536,0.004795555439260271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,6144,0.22631732622782388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,16384,0.03477955526775784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,16384,0.023390221926901076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,4096,0.01145600030819575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,7168,0.010935111178292168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,8192,1.9299466874864366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,32,0.0046568889584806235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,2048,0.009463110731707679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,12288,0.025814221964942083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,65536,0.07870578103595309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,8192,0.007897777689827813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,10240,0.025943999489148457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,16384,0.8797911008199056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,64,0.006186666587988536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,6144,0.009930666950013902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,6144,0.011323555476135679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,512,0.015191111299726697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,128,0.0030177777840031516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,4096,0.012462221913867526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,12288,0.012672888735930124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,5120,0.1359928846359253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,64,0.0037831109431054858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,3584,0.045089777972963124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,8192,0.11134044329325359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,64,0.004953777624501122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,7168,0.029231998655531142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,5120,0.09608178006278144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,8192,0.026386666629049513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,1536,0.035563555028703474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,1536,0.008301333420806462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,5120,0.009342222577995723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,2048,0.10635733604431152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,256,0.010295111272070143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,16384,0.7296542061699761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,3072,0.010678222609890832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,7168,0.01331555595000585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,8192,0.12409333388010661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,8192,0.031154665682050917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,65536,5.612628512912327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,32,0.007928888830873702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,8192,0.01516711049609714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,512,0.004638222356637319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,64,0.0026817777090602448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,65536,0.08775822321573894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,12288,0.46782223383585614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,32,0.00868088917599784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,512,0.05769422319200304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,3584,0.338135110007392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,512,0.00889155517021815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,6144,0.018557333283954196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,2048,0.011774222056070963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,128,0.0030373332815037835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,768,0.005467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,32,0.00564444437623024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,8192,0.018522666560279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,512,0.005792888916201062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,10240,0.14154222276475695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,128,0.0031911111954185697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,512,0.02537511123551263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,4096,0.019489778412712943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,128,0.0063528890411059065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,3584,0.026017778449588354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,768,0.019959999455346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,64,0.005078222188684675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,16384,0.01960177719593048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,512,0.0032791110376516977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,64,0.002672000063790215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,2560,0.010098666780524785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,7168,0.012344000240166983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,10240,0.17570932706197104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,32,0.007898666792445714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,2560,0.007105777661005656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,1536,0.03425511055522495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,2048,0.01643822259373135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,128,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,2560,0.26796266767713756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,32,0.01314133322901196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,256,0.011934222446547614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,16384,0.017649778061442904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,64,0.006770666688680649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,256,0.011037333144081963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,128,0.004078222231732475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,512,0.16404710875617132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,65536,0.07602222098244561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,768,0.003705777641799715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,1536,0.005030222237110138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,7168,0.011375111010339526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,7168,0.02625511089960734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,512,0.006835555450783835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,10240,0.023016000787417095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,5120,0.037787556648254395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,16384,0.09027910894817776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,8192,0.06410221921073066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,1536,0.06667555703057183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,768,0.0035377778112888336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,256,0.015819556183285184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,768,0.0053688887920644544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,768,0.003683555457327101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,10240,0.0248391115003162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,512,0.00480000012450748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,7168,0.056341330210367836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,12288,0.054495112763510815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,1536,0.011759111450778114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,8192,0.021627555290857952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,10240,0.016538666354285348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,128,0.0068986668354935115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,16384,0.07037066751056247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,6144,0.13224977917141384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,64,0.005051555732885997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,2048,0.012103999654452005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,3584,0.025460445218616064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,7168,0.02493155499299367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,768,0.003917333152559069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,1024,0.032772445016437106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,12288,0.09591822491751777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,64,0.0026266665922270883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,16384,0.04776800009939405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,32,0.012934221989578672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,1536,0.011233777635627322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,4096,0.25164445241292316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,3584,0.015273777974976433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,2048,0.006721777634488211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,1536,0.007840889195601145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,1024,0.022275555464956496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,2560,0.028524445162879095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,768,0.0057964444988303715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,65536,0.20257955127292207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,2560,0.00598933340774642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,6144,0.01593777702914344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,1024,0.007904888855086433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,768,0.005913777897755305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,8192,0.009591999981138442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,256,0.010904889139864178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,512,0.004769777672158347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,1536,0.005720888988839255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,32,0.004223111189073986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,768,0.02490844494766659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,256,0.003240888938307762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,64,0.0028213332924577924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,2560,0.007869333028793335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,3072,0.0155102229780621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,2560,0.022482666704389784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,7168,0.02032533288002014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,65536,0.1921457714504666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,10240,0.07958577738867866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,65536,0.33243732982211643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,1024,0.006627555522653792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,12288,0.006441777778996362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,65536,1.486425823635525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,65536,0.09067289034525554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,2048,0.005873777800136142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,16384,0.05885422229766846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,1024,0.005156444592608346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,1024,0.02125688890616099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,2560,0.012991111311647626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,2048,0.005974222388532426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,768,0.01202133297920227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,3584,0.010055999788973067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,65536,0.5529493225945367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,3072,0.05459910962316725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,10240,0.018200000127156574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,128,0.006666666517655055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,2048,0.016985777351591323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,32,0.003010666618744532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,7168,0.014595554934607612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,3584,0.03938755393028259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,6144,0.29781421025594074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,768,0.022830223043759663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,64,0.007315555380450354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,1024,0.01681066718366411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,65536,0.5473706457349989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,7168,0.01849688920709822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,4096,0.03222844335767958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,4096,0.03013777732849121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,3584,0.012058667010731168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,8192,0.0069013333155049225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,2560,0.018485332528750103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,16384,0.02022133270899455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,512,0.02622311148378584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,3072,0.010122666756312052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,64,0.005940444353553984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,3584,0.021597334080272253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,3584,0.00905777762333552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,32,0.0026968889352348116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,256,0.003519999898142285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,6144,0.024446222517225478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,512,0.005428444594144821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,32,0.12703643904791936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,2560,0.03036266565322876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,32,0.0032684444967243406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,3584,0.014360000689824423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,2560,0.007126222054163615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,1536,0.0046977777447965415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,12288,0.19402666886647543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,256,0.003079111170437601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,1536,0.006692444284756978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,3584,0.01592177814907498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,3584,0.00684799998998642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,2560,0.007146666447321574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,7168,0.03477955526775784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,65536,0.06803644365734525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,65536,0.155166228612264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,128,0.004605333424276776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,2048,0.018437332577175565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,65536,4.578065660264757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,8192,0.1339555581410726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,4096,0.18250132931603325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,6144,0.017285333739386666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,1024,0.005316444569163852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,16384,0.02314044369591607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,32,0.0028151110228565005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,512,0.005583999885453119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,5120,0.15235643916659886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,1024,0.005045333256324132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,3584,0.04624799887339274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,64,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,6144,0.049116445912255176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,5120,0.27052532302008736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,12288,0.0441964434252845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,2560,0.029096888171301946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,768,0.008223111430803934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,3072,0.2704586717817518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,2560,0.009169778062237633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,65536,0.11626311143239339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,65536,2.852773242526584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,512,0.0035955554081334006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,5120,0.029309332370758057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,1536,0.004761777818202972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,2560,0.02955377764172024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,128,0.004483555754025777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,512,0.005884444548024072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,16384,0.025723555021815833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,1536,0.014343111051453484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,768,0.004741333425045013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,1024,0.008916444248623317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,6144,0.42294934060838485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,1536,0.04572622312439812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,128,0.006917333437336816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,3584,0.017982222967677645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,6144,0.029198222690158423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,10240,0.017345777816242643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,4096,0.010399999717871347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,3584,0.009783999787436591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,2560,0.007679999702506595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,10240,0.03701600101259019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,64,0.002543111021320025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,2560,0.006695999867386288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,2048,0.006708444406588872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,256,0.002963555562827322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,4096,0.22252533170912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,1536,0.004620444443490771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,3584,0.02601066728432973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,5120,0.016294222739007738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,10240,0.019727110862731934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,1536,0.022407111194398668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,32,0.0025333333760499954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,768,0.030379556947284277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,2048,0.0064791109826829694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,512,0.0032924444725116095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,3072,0.008044444852405125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,3584,0.1861928833855523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,256,0.003014222201373842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,7168,0.019827556278970506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,4096,0.018442667192882962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,16384,0.01907911068863339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,2560,0.008813333180215623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,8192,0.03993511199951172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,128,0.006641777853171031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,3072,0.02574311031235589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,65536,0.035813334915373064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,4096,0.012103111379676394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,12288,0.33213867081536186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,768,0.011391999820868174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,512,0.013663999736309052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,6144,0.010449777874681683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,16384,0.020267556111017864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,256,0.017103110750516255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,12288,2.8692150115966797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,3072,0.013717333475748697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,5120,0.016568889220555622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,16384,0.0514355566766527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,768,0.029042667812771265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,65536,0.06732177734375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,1024,0.006015111174848344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,65536,0.06715199682447645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,65536,0.36432978841993546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,10240,0.08214666446050008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,12288,0.017837334010336135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,256,0.003097777772280905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,12288,0.020604444874657523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,10240,0.01664266652531094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,3072,0.11890933248731826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,3072,0.1735502216551039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,12288,0.06390221913655598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,6144,0.23313511742485893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,768,0.005606222069925732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,128,0.0027991111079851785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,16384,0.22197333971659342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,64,0.03566044569015503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,128,0.004148444367779626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,4096,0.017163554827372234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,2048,0.005208888815508948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,256,0.004256000121434529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,768,0.0064177778032090925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,2048,0.006591111007663939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,128,0.005443555613358815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,3072,0.007963555554548899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,1536,0.012439111040698158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,1536,0.004610666798220741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,10240,0.03420177764362759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,1024,0.012989333106411828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,1024,0.0058453331391016645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,16384,0.021925333473417494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,256,0.0034168888297345904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,64,0.0026186667382717133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,768,0.003576888806290097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,8192,0.01036088913679123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,32,0.00847466703918245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,128,0.007064888874689738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,128,0.011699555648697747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,1024,0.011050666371981302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,3072,0.008664888640244802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,512,0.0046640001237392426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,2560,0.010301333334710863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,5120,0.01773422294192844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,5120,0.01316177762216992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,6144,0.02402399977048238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,2048,0.006721777634488211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,768,0.010886222124099731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,1024,0.0057653333577844836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,128,0.0029813332690132987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,2560,4.18288336859809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,6144,0.013400000002649097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,512,0.003526222374704149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,256,0.0058355554938316345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,65536,0.3212471008300781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,2048,0.016234666109085083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,4096,0.0542702211274041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,3072,0.012151999606026543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,32,0.005424889011515512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,512,0.005807111246718301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,1024,0.005210666606823604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,10240,0.03705866800414191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,768,0.033839109871122576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,16384,0.024085332949956257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,1536,0.01051288925939136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,12288,0.05392711030112373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,16384,1.5917883978949652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,5120,0.008512000242869059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,256,0.006882666713661618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,16384,0.034641779131359525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,7168,0.019813333948453266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,256,0.009025777379671732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,4096,0.01126933346192042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,32,0.009936889012654623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,256,0.0054560001525614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,2560,1.0110692977905273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,4096,0.027093332674768236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,2048,0.012783110969596438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,768,0.008657777474986183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,64,0.006222222414281633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,64,0.020840888222058613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,65536,4.7736735873752165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,768,0.009550222092204625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,10240,0.018187556001875136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,256,0.004559111264016894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,1536,0.013771555489963956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,5120,0.030072887738545735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,3584,0.008638222184446123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,2048,0.019211555520693462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,2560,0.012512889173295764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,2560,0.01702933344576094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,16384,0.02885599931081136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,7168,0.04678577846950955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,3584,0.023079999619060095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,128,0.005567111074924469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,1024,0.007857778006129796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,128,0.005222222043408289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,10240,0.01367111090156767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,3072,0.014693333042992486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,16384,0.051784889565573804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,64,0.004991111242108875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,8192,0.022283555732833013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,8192,0.012773333324326409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,4096,0.00775911079512702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,65536,0.07297244336869982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,8192,0.2891217867533366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,512,0.009664889011118147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,6144,0.1892684433195326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,12288,0.04767733481195238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,8192,0.015769778026474845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,128,0.004387555436955558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,8192,0.006506666541099548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,1024,0.006247111078765657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,4096,0.2240062289767795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,3584,0.007505777809354994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,3072,0.08719822433259751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,2560,1.0275804731580946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,10240,0.008070222205585903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,5120,0.013440888788965015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,65536,0.07759555843141344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,256,0.003918222255176968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,256,0.004294222427739037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,2560,0.01477688882086012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,7168,0.03431111243036058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,10240,4.396535237630208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,1536,0.008192888564533657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,64,0.0032622222271230486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,256,0.02648888859483931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,2048,0.005416888743638992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,8192,0.00813333359029558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,32,0.01518311103185018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,7168,0.025961778230137292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,8192,0.015861334072219003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,2560,0.010052444206343757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,32,0.007749333149856991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,512,0.006870222174459034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,7168,0.01274311128589842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,4096,0.009272000028027428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,2560,0.01458666721979777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,1024,0.02218755582968394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,768,0.008962666822804345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,3584,0.009688888986905416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,64,0.008060444560315874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,5120,0.0143857780430052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,8192,1.1650009155273438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,32,0.00638488887084855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,5120,0.1458266708585951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,3584,0.03879377908176846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,2048,0.01757866640885671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,32,0.018292443619834054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,2048,0.016172443826993305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,32,0.005895111295912001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,3584,0.007905777957704332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,5120,0.01686311099264357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,16384,0.01146488885084788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,5120,0.03368799885114034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,65536,0.19272622797224256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,8192,0.1529946724573771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,65536,0.13080444600847033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,65536,0.2854417694939507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,32,0.008760888543393876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,1024,0.01347555551264021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,32,0.006041777630647023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,12288,0.021381333470344543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,512,0.07179555628034803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,3072,0.02740977704524994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,6144,0.21821422047085234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,65536,0.06810222069422404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,256,0.0031662223239739737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,1536,0.004776000148720211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,4096,0.0358533329433865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,1536,0.020642666353119742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,2048,0.0061928890645504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,10240,0.029233776860766943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,7168,0.010371555884679159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,16384,0.09687822394900852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,128,0.004407111141416761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,512,0.007276444799370236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,2048,0.006037333359320958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,64,0.005316444569163852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,512,0.01815200017558204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,32,0.0880062248971727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,512,0.004902222090297275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,4096,0.027985778119828966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,16384,0.03944177760018243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,2048,0.016745777593718637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,3584,0.012511110968059964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,512,0.0034959999223550162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,512,0.004776000148720211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,256,0.004644444419278039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,512,0.0037600000699361167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,7168,0.026564444104830425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,10240,0.018231999542978074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,8192,0.011426666544543372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,8192,1.4174053404066298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,1536,0.0061048890153567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,512,0.00869066682126787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,3072,0.00903733323017756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,65536,19.928697374131946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,3072,0.020624889267815482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,256,0.005471111171775394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,10240,0.06672444608476427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,5120,0.01332711097266939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,8192,0.01648977730009291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,5120,0.19344000021616617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,4096,0.15790043936835393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,65536,0.06683378087149726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,7168,0.055996444490220815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,3072,0.014491554763582019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,128,0.009148444566461775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,8192,0.1371262205971612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,1536,0.00944622192117903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,7168,0.022776888476477727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,10240,0.06920266813702054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,3072,0.011152889165613385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,4096,0.008088000118732452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,12288,1.1865706973605685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,3584,0.01000088867213991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,1536,0.010096000300513374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,65536,0.3274293210771349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,4096,0.019448889626397025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,1024,0.004026666697528628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,64,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,256,0.003107555624511507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,1024,0.016597333881590102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,65536,0.34356265597873265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,8192,0.03305066625277201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,32,0.005505777895450592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,3072,0.010679110884666443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,2560,0.011418666276666852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,128,0.005052444421582752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,7168,0.010816000401973724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,3584,0.018599111172888014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,7168,0.39695199330647785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,10240,0.3433066738976373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,12288,2.5681671566433377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,2560,0.011516444385051727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,7168,0.09652977519565159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,65536,0.06758222315046522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,3584,0.016977777083714802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,3584,0.007892444729804993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,2560,0.012496888637542725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,1536,0.010231999887360467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,12288,0.042061332199308604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,2560,0.01535999940501319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,256,0.0047733332547876565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,2560,0.20490754975212944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,65536,0.3929591178894043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,6144,0.013391111459996967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,1024,0.004020444634887907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,1024,0.01698844465944502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,65536,0.19139911068810356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,12288,0.06396266486909655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,512,0.011306666665607028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,4096,0.08877866797977024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,5120,0.016379555066426594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,7168,0.025211556090248957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,7168,0.02608799934387207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,7168,0.020584889584117465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,3584,0.00816533300611708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,2560,0.030908445517222088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,512,0.007085333267847697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,3072,0.007676444119877285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,512,0.004374222209056218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,128,0.004337777694066365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,2048,0.006252444452709622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,1024,0.008414222134484185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,128,0.00612088872326745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,10240,0.017528888252046373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,16384,0.08697244193818833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,5120,0.006591111007663939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,256,0.0037191112836201987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,768,0.006213333457708359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,256,0.003072000005178981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,16384,0.02128355536195967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,5120,0.01256711118751102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,2560,0.005598222215970357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,65536,1.6245732837253148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,8192,0.10417422321107651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,7168,0.014999111493428549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,3584,0.15895288520389134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,256,0.26536620987786186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,768,0.005407111098368962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,1024,0.0069022224181228215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,512,0.007428444094128079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,16384,0.01867999964290195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,3072,0.058931556012895375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,3584,0.009749333063761393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,8192,0.032991111278533936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,512,0.005407111098368962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,256,0.004238222208287981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,2560,0.016166221764352586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,1024,0.020047111643685233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,16384,0.05595733059777153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,6144,0.011913778053389655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,256,0.0031902222997612427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,4096,0.08324355549282499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,65536,0.06491288873884413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,16384,0.25759021441141766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,6144,0.011225777367750803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,12288,0.10121333599090576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,4096,0.010299555957317352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,7168,0.03554577628771464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,32,0.0028071111689011254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,32,0.005119110975000593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,32,0.0033760000434186724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,1536,0.009283555878533257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,1536,0.005408000200986862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,2048,0.02405066622628106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,256,0.003144888828198115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,2560,0.01761777698993683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,6144,0.08007466793060303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,6144,0.010101333260536194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,128,0.00701333334048589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,65536,0.06915377908282809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,16384,0.03408088948991563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,512,0.006008000009589725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,7168,0.07336355580223931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,2048,0.010785777535703448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,2560,0.010915555887752108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,2048,0.011725333001878528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,3072,0.027772444817754958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,1024,0.011739555332395764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,768,0.00942311104800966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,3584,0.007440889047251807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,768,0.006043555421961679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,3584,0.01071911139620675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,5120,0.01145155562294854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,2560,0.10343022478951348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,6144,0.013253333667914072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,2048,0.06111822525660197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,65536,0.8297964202033149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,64,0.003440888805521859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,5120,0.014713777436150445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,12288,0.020119110743204754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,3584,0.01862844493654039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,256,0.003626666549179289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,65536,0.10461333062913682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,32,0.0027235555979940626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,5120,0.0092622223827574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,7168,0.02348711093266805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,3584,0.00908088849650489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,1536,0.006494222415818109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,1536,0.008952000074916417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,1536,0.02946044339074029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,768,0.042240000433391996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,16384,0.04907911022504171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,1024,0.0047048889100551605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,10240,0.21325511402553984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,4096,0.01997066703107622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,12288,0.011365333365069495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,3584,0.015583111180199517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,512,0.11053599913914998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,8192,0.01493866741657257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,3584,0.006766222417354584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,64,0.002625777696569761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,64,0.0027537778433826235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,12288,0.02427022159099579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,10240,0.04710666669739617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,768,0.008555555509196388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,2048,0.008970666262838576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,6144,0.04526666800181071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,2048,0.012576888832781049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,8192,0.060565332571665444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,6144,0.030456887351142034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,32,0.03861600160598755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,64,0.00258577780591117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,768,0.021592888567182753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,3584,0.020503110355801053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,2560,0.009415110780133141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,3072,0.008767999708652496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,8192,0.15935199790530735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,4096,0.008428444465001425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,65536,0.09975288973914252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,10240,0.028324445088704426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,128,0.005057777795526716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,4096,0.016711110870043438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,3072,0.0096124443742964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,64,0.0030168888883458245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,1024,0.006109333286682765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,8192,0.020022221737437777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,65536,0.1616604460610284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,5120,0.05096888873312208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,1536,0.0063591111037466265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,7168,0.014564444621404013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,1536,0.005642666584915585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,10240,0.01407199932469262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,7168,0.010493333141009012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,1024,0.03520444366667006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,1536,0.006798222247097228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,12288,0.02963288956218296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,7168,0.5013590918646919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,512,0.008627555436558193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,768,0.01812000075976054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,1536,0.009591999981138442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,256,0.06813600328233507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,8192,0.025776000486479864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,768,0.008958222137557136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,512,0.010776888993051318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,4096,0.033768001529905535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,7168,0.6391653484768337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,32,0.005966222120655908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,10240,0.03821511069933573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,5120,0.09018577469719781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,1536,0.01458666721979777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,7168,0.012124444047609964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,3584,0.06227111154132419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,3584,0.023588443795839947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,6144,0.010091555615266165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,12288,0.373257769478692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,65536,0.23708976639641655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,3584,0.008550222549173567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,64,0.0026293332791990707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,512,0.003481777591837777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,2560,0.08064177963468763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,10240,0.03844177722930908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,768,0.006760889043410619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,512,0.005716444303592046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,1536,0.0074142225914531285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,256,0.007701333198282454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,65536,0.20249244901869032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,7168,0.007865777446164025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,2048,0.006068444500366847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,128,0.003738666574160258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,1024,0.012216000093354119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,1024,0.006005333529578314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,3072,0.007259555160999298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,2560,0.016028443972269695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,1024,0.011588444312413534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,3584,0.022583999567561682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,3072,0.013260444833172692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,2560,0.007335111498832703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,1024,0.039988444911109075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,256,0.004471999903519948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,128,0.012994666894276937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,6144,0.5624000231424967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,65536,0.290279123518202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,12288,0.009418666362762451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,8192,0.04007110993067423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,128,0.006603555546866522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,128,0.0041928887367248535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,1024,0.00887822194231881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,4096,0.043172445562150746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,5120,0.016171556380059984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,5120,0.009759999811649323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,7168,0.010667555862002902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,256,0.0047351109484831495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,4096,0.45620001686943906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,7168,0.007763555480374231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,10240,0.04261777798334757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,10240,1.0474150975545247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,7168,0.015341333217091031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,7168,0.09652444389131333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,1536,0.013400000002649097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,6144,0.019926221834288705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,64,0.0027457777824666765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,16384,0.014613333675596448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,6144,0.01758311192194621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,1024,0.07498933209313287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,1536,0.007785778078767989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,32,0.005583999885453119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,128,0.1076800028483073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,3584,0.023307555251651343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,128,0.005292444593376583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,6144,0.026593777868482802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,3584,0.06210222509172228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,2560,0.011948444777064852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,64,0.0026319999661710528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,65536,0.1051902241177029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,6144,0.024528889192475214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,12288,0.04267200165324741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,32,0.011510222322411008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,4096,0.04705244302749634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,2048,0.06483378012975057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,12288,0.09112799829906887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,512,0.10009511311848958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,7168,0.015291555060280694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,64,0.0069217777086628815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,256,0.01092177795039283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,10240,0.024329778220918443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,7168,0.014670222997665405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,5120,0.014955555399258932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,768,0.011614222493436603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,12288,0.058189332485198975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,2048,0.04356000158521864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,16384,0.09284444650014241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,768,0.19508443938361275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,768,0.015434667468070984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,64,0.006565333240562015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,256,0.004926222066084544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,8192,0.013735999663670858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,3072,0.03861866725815667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,12288,0.030368000268936157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,7168,0.027958220905727808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,5120,0.04361688759591845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,8192,0.04010755485958523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,65536,0.2418959935506185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,512,0.004655111167165968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,3072,0.030290666553709242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,6144,0.11126222875383164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,4096,0.012846222354306115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,64,0.004401777767472797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,256,0.004309333446953031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,256,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,256,0.009853333234786987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,2560,0.011993777420785693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,64,0.005108444641033809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,5120,0.2559911145104302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,8192,0.10499733024173313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,3072,0.5843608644273546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,4096,0.07585155301623873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,32,0.004481777962711123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,12288,0.22958400514390734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,8192,0.022307554880777996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,3072,0.009827555881606208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,16384,6.649628533257379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,3072,0.05109066764513651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,128,0.005021333280536863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,3584,0.007655999726719326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,16384,0.1463973389731513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,16384,0.030664887693193223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,7168,0.0798542234632704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,128,0.00849511143234041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,10240,0.013399110900031196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,16384,0.026570666167471144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,768,0.0047395556337303585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,12288,0.04062311185730828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,12288,0.03935822182231479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,65536,0.21145689487457275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,6144,0.007232888705200619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,16384,0.07376711236106025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,2048,0.006781333436568578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,768,0.006216889040337668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,1536,0.0057955553962124725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,8192,0.17645866341061064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,6144,0.013801777528391944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,6144,0.10610577795240615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,32,0.006684444430801604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,5120,0.007243555453088548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,7168,0.017825777331988018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,10240,0.022288888692855835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,1024,0.005807111246718301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,5120,0.02369155486424764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,2048,0.010255110760529837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,6144,0.022008889251285132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,1024,0.009727111293209923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,4096,0.010658666491508484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,6144,0.01128888875246048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,1536,0.006544000158707301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,4096,0.018626666731304593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,5120,0.02423733307255639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,6144,0.00961688905954361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,768,0.0057608890864584185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,7168,0.03939911060863071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,1536,0.006974221931563483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,1024,0.0885288914044698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,1024,0.01663555536005232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,12288,0.06049333016077677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,64,0.007183111376232571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,3072,0.0081395556529363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,1536,0.08526400062772964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,5120,0.010144000252087912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,16384,0.021137777302000258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,5120,0.00852711084816191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,512,0.00941422250535753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,1024,0.006025777922736273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,128,0.0041653331783082755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,7168,1.38573424021403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,512,0.01370666672786077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,3072,0.009450666606426239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,768,0.006516444600290722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,1024,0.01202044470442666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,65536,0.23493422402275932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,768,0.005143110950787862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,12288,0.03420089019669427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,2560,0.007130666739410824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,768,0.0057911111248864075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,8192,0.35399911138746476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,1536,0.011352889239788055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,32,0.0034684443639384378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,5120,0.3990933365292019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,256,0.0033582221302721235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,32,0.005515555540720622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,32,0.003137777869900068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,32,0.006696888970004187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,64,0.00645955569214291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,512,0.008225777910815345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,512,0.024136889312002394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,512,0.003910222401221593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,10240,0.016024000114864774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,64,0.00676711110605134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,32,0.003104888937539525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,6144,0.006732444382376141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,10240,0.04758844441837735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,1536,0.004413333204057482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,512,0.003181333343187968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,3072,0.07681244611740112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,7168,0.05903910928302341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,3584,0.029393778906928167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,2048,0.009170666337013245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,64,0.008953777452309927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,3584,0.006705777926577463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,7168,0.01811644434928894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,1536,0.011692444483439127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,1536,0.006615110983451207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,2560,0.02324177821477254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,16384,0.05947022305594551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,256,0.0035573335157500375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,2560,0.007272889216740926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,2560,0.006253333141406377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,768,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,1024,0.006317333214812809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,64,0.0063662222690052455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,16384,0.07283911440107557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,3072,0.014546665880415173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,10240,0.016664889123704698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,2048,0.010537777509954242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,16384,0.12395289209153916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,2048,0.005388444496525659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,10240,0.017478222648302715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,65536,0.24109334415859648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,256,0.00370755543311437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,12288,0.06441155407163832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,64,0.003735110991530948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,1024,0.004035555654101902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,10240,0.027448000179396734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,16384,0.15689867072635227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,65536,0.08768177694744533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,512,0.005758222192525864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,12288,0.026758222116364375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,2560,0.033722665574815534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,3072,0.059699555238087974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,128,0.006515555497672822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,1536,0.011167111496130625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,5120,0.014282666974597506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,6144,0.40763645701938206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,32,0.005395555661784277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,10240,0.7415324317084419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,256,0.00480000012450748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,10240,0.054472890165117055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,1536,0.015840889679061044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,8192,0.015222221612930298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,256,0.007460444337791867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,4096,0.006671999891599019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,7168,0.03177155719863044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,128,0.005286222116814719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,512,0.005134222408135732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,5120,0.09880177842246161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,8192,0.013742222554153867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,256,0.003012444410059187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,10240,0.03913244605064392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,3584,0.018760888112915885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,256,0.04471822248564827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,128,0.00432977784011099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,8192,0.18222222063276503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,32,0.004371555729044808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,3584,0.011014222270912595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,12288,0.025015999873479206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,4096,0.018738667170206707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,768,0.0047484445903036325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,3072,0.1515902280807495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,128,0.005418666534953647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,1536,0.0365866654449039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,512,0.005655111124118169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,2560,0.018704000446531508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,16384,0.20024443997277153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,64,0.005493333356248007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,65536,0.40545066197713214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,8192,0.01347555551264021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,3072,0.013532444834709167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,3072,0.03561244408289591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,65536,0.0697644419140286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,1536,0.009687999884287516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,3584,0.01938666734430525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,1536,0.03943377733230591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,64,0.005175999883148405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,2560,0.006258666515350342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,7168,0.055233776569366455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,5120,0.021314667330847845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,8192,0.019129777948061626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,5120,0.020498666498396132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,8192,0.02392888896995121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,2048,0.02966844373279148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,1536,0.004649777793222004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,4096,0.02958400050799052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,1024,0.01147822207874722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,10240,0.04702577657169766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,1024,0.03951199849446615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,12288,0.018516444497638278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,16384,0.014467555615637036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,7168,0.02844088938501146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,3584,0.008541333178679148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,3584,0.007106666763623555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,5120,0.03418577710787455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,2048,0.013937777943081327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,128,0.01167555567291048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,128,0.0026462222966882917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,2048,0.006133333262470033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,64,0.005321777943107817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,512,0.0046728890803125174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,16384,0.019879110985332064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,4096,0.014901333385043673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,2048,0.007507555186748505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,1536,0.022571555442280237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,2560,0.007666666474607255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,12288,0.144923554526435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,3584,0.007556444240940942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,2560,0.012871111432711283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,3584,0.032186667124430336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,2560,0.0070471109615431885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,2560,0.09045777718226115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,512,0.03586488962173462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,16384,0.8094888793097602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,256,0.003338666632771492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,3584,0.007347555624114142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,2048,0.01109866632355584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,768,0.005593777944644292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,4096,0.010520888699425591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,1536,0.006653333289755716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,128,0.00535733335547977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,256,0.00453955555955569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,7168,0.03717866539955139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,2048,0.007358222372002072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,16384,0.06274666388829549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,3584,0.023648000425762598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,32,0.0027520000520679685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,3584,0.0703502231174045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,8192,0.006182222316662471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,16384,0.02353511088424259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,3584,0.13773955239189997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,65536,0.09212622377607559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,16384,0.031734221511416964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,5120,0.011904888682895236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,1536,0.14100533061557347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,3072,0.010479110810491772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,768,0.005858666780922148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,2048,0.04020177655749851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,768,0.015941333439615037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,8192,0.27633688184950084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,1536,0.005701333284378052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,5120,0.029990222718980577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,3072,0.09976889027489556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,65536,0.14717155032687718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,3072,0.017106667160987854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,12288,0.03815999958250258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,16384,1.2193226284450955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,768,0.015004444453451367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,12288,0.019561777512232464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,512,0.004039999925427967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,5120,0.02310666607485877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,3072,0.009296000003814697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,65536,0.009451555709044138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,3072,0.008734222087595198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,7168,0.06849510802163018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,2560,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,256,0.0033768889390759995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,768,0.0036311112344264984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,128,0.0048071112897661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,65536,0.06811377737257215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,32,0.0053075556125905775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,65536,0.33400623003641766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,512,0.004084444294373195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,3072,0.010199111368921068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,768,0.004036444342798657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,768,0.020785777105225455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,32,0.3553066783481174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,12288,0.372195561726888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,10240,0.008023999631404877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,5120,0.04652888907326592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,6144,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,3584,0.014608888162506951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,5120,0.33348798751831055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,8192,0.010279999838935005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,10240,0.020402666595247056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,10240,0.0277502222193612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,768,0.10720000002119277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,3072,0.018529777725537617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,32,0.003617777592606015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,4096,0.02903999884923299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,1024,0.005703111075692707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,4096,0.017419555120997958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,4096,0.014170666535695394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,64,0.00526577772365676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,512,0.0038568890757030914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,16384,0.036320888333850436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,10240,0.01538222200340695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,768,0.005129777722888523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,12288,0.024872000018755596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,16384,0.3346542252434625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,8192,0.0251555558707979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,1024,0.009259555902745988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,4096,0.0170168893204795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,2560,0.008463111188676622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,64,0.008648888932334052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,32,0.0052675555149714155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,512,0.016935999194780987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,12288,0.04478222131729126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,3072,0.023244443866941664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,256,0.006717333363162146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,4096,0.019690665933820937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,3072,1.2546240488688152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,4096,0.013983999689420065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,10240,0.0723866687880622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,768,0.005068444543414646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,7168,0.05492266681459215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,5120,0.009425777528021071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,768,0.0041511112617121804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,3072,0.06103377872043186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,4096,0.042048000627093844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,4096,0.027909331851535376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,3072,0.15578667322794595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,5120,0.017179555363125272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,6144,0.08209866947597928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,12288,0.2043315569559733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,128,0.003920888735188378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,8192,0.006568000134494569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,2048,0.011865778101815118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,65536,0.08114311430189344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,7168,0.006183111005359226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,10240,0.016964443855815463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,16384,0.019440889358520508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,12288,0.01850311126973894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,768,0.00553955551650789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,1536,0.005500444521506627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,4096,0.022914666268560622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,1024,0.0073520003093613525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,3584,0.007483555210961237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,1536,0.008185777399275038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,1024,0.005633777628342311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,65536,0.09878222147623698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,2048,0.010936888555685679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,65536,0.18526132901509604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,128,0.005423999908897612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,2048,0.00775911079512702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,4096,0.012531555361217923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,7168,0.008873778084913889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,3072,0.007146666447321574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,3072,0.015218666858143277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,8192,0.021261332763565913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,2048,0.006587555425034628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,64,0.04938044481807285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,1536,0.018138666947682697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,3584,0.027296889159414504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,512,0.0033484444850020935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,256,0.021710221966107685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,5120,0.013111111190583972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,12288,0.11692355738745795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,4096,0.026660444007979497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,65536,0.23917955822414824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,3584,0.017747556169827778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,12288,0.03132889005872939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,1024,0.005531555662552516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,8192,0.01706133286158244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,1536,0.021958221991856892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,2560,0.008534222013420528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,256,0.0041866666740841335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,128,0.00425155543618732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,1536,0.005852444304360284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,16384,1.1721021864149306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,12288,0.033477332856920027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,3584,0.1013608906004164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,256,0.0035297779573334586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,4096,0.025736888249715168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,3072,0.012293332980738746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,16384,0.011151110960377587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,7168,0.387264887491862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,3584,0.006593777901596493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,4096,0.12715289327833387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,12288,0.11843378014034694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,512,0.005177777674463061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,12288,0.02737244466940562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,2560,0.006098666538794835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,3072,0.006728000111050076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,7168,0.019111999207072787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,10240,0.11767289373609756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,512,0.11839289135403103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,6144,0.036068442795011736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,8192,0.01276088919904497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,65536,0.07572978072696261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,128,0.01575644479857551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,65536,0.09024266401926677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,768,0.004855111241340637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,10240,0.12114044030507405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,512,0.005697777701748743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,5120,0.017310221989949543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,16384,0.7169093555874295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,64,0.0034773333205117118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,512,0.005159111072619756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,512,0.0058604445722368025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,16384,0.16139554977416992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,512,0.011740444435013665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,8192,0.006121777825885349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,2048,0.03616444600952996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,1536,0.050069332122802734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,16384,0.021382222572962444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,64,0.015391111373901367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,12288,0.04295911060439216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,8192,0.011504000259770287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,4096,0.017416000366210938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,6144,0.021292444732454088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,128,0.00812977800766627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,6144,2.511209699842665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,10240,0.028883556524912517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,3584,0.011409777734014722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,3072,0.007995555798212687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,128,0.003469333259595765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,4096,0.01443288889196184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,256,0.005049777941571342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,256,0.005579555614127054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,2048,0.008080888953473832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,32,0.0025555555605226094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,64,0.005020444591840108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,65536,0.18894844584994844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,1536,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,512,0.0033822223130199644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,2048,0.011419555379284753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,2048,0.006041777630647023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,4096,0.020480889413091872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,8192,0.016404444972674053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,768,0.004259555704063839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,64,0.0027173333283927706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,512,0.015605333778593274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,3584,0.14174400435553655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,1024,0.03881422347492642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,65536,0.06660711103015475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,128,0.004456888884305954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,7168,0.05034844411744011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,2048,0.0052915554907586836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,1536,0.00638755535085996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,3072,0.007359999749395583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,2560,0.0075733330514695905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,8192,0.26500267452663845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,10240,0.026122666067547266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,2048,0.020076443751653034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,8192,0.304511997434828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,12288,0.03889244463708665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,256,0.011337777806652917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,1024,0.008632889224423302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,3072,0.008372444245550368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,12288,0.037061333656311035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,512,0.0047066667013698155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,2048,0.012630222572220696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,12288,0.02457155618402693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,16384,0.026537777649031744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,65536,0.21984710958268908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,4096,0.007891555627187094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,2560,0.007361777954631382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,2560,0.005971555494599872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,5120,0.021553777986102637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,7168,0.0266311119000117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,3584,0.013169777890046438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,4096,0.009472000102202097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,128,0.02299822204642826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,3584,0.012186666329701742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,8192,0.03052088949415419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,12288,0.04549955659442478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,768,0.015209777487648858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,64,0.006725333217117522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,3584,0.011115555961926779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,2560,0.007166222565703922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,65536,0.24067555533515084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,8192,0.024346666203604803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,2048,0.07974221971299914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,12288,0.017667555146747164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,65536,0.18479555182986787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,512,0.013204444613721637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,128,0.0033777778347333274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,1024,0.0041511112617121804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,1024,0.008789333204428354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,1024,0.005502222312821283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,512,0.02529866662290361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,10240,0.01644177734851837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,768,0.0058746664888328975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,32,0.00851822230550978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,8192,0.022076444493399724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,6144,0.011766222616036734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,5120,0.0795875522825453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,3584,0.2886062198215061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,65536,0.21179555522070992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,8192,0.10783288876215617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,2048,0.019183999962276883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,1536,0.009323555562231276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,1536,0.011810666984981961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,5120,0.009487110707494948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,1024,0.0041448887851503165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,2048,0.005476444545719359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,256,0.003649777836269803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,3072,0.0075697774688402815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,4096,0.009220444493823582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,16384,0.7957848972744412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,10240,0.01978488928741879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,768,0.00906044410334693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,1024,0.007893333004580604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,5120,0.13548089398278132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,6144,0.024080889092551336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,5120,0.010503110786279043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,3584,0.019034667147530448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,64,0.0053288886944452924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,6144,0.03397688931889004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,3584,0.0564933353000217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,3072,0.00981422182586458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,65536,0.12275644143422444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,7168,0.02234399980968899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,8192,0.44121776686774355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,12288,0.09299911393059625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,65536,0.13116266992357042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,32,0.00263288886182838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,3584,0.006768888897365994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,10240,0.31627021895514595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,6144,0.014090667168299357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,64,0.005400889035728242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,64,0.09589955541822646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,256,0.006481777876615524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,3584,0.021759111020300124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,65536,0.31077777014838326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,3072,0.02641422218746609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,4096,0.00831200016869439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,65536,0.15124089188045925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,8192,0.01516355574131012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,3072,0.008253333469231924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,128,0.0028497777465316984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,2048,0.011376889215575324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,16384,0.04929244518280029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,6144,0.010409778190983666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,65536,1.218897819519043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,64,0.003702222059170405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,7168,0.01698666645420922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,16384,0.02474044428931342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,4096,0.5351777606540257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,768,0.00963288876745436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,16384,0.020351111888885498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,3584,0.011750222080283694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,1536,0.006213333457708359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,2048,0.020340444313155282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,64,0.005970666805903117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,7168,0.010382222632567087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,3584,0.00944088896115621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,3072,0.2747013303968641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,6144,0.045985778172810875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,5120,0.019187556372748483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,32,0.026169778572188482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,1024,0.003989333493842019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,32,0.004341333276695675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,512,0.018495111001862418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,16384,0.058437334166632764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,16384,0.03721866673893399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,7168,0.059783114327324756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,512,0.003370666669474708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,65536,0.4104471206665039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,64,0.012984000146389008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,512,0.004853333450025982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,7168,0.027907556957668726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,1536,0.006309333360857434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,5120,0.013293333351612091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,16384,0.05486666825082567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,10240,0.026706667410002813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,65536,1.0724764929877388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,128,0.00480355570713679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,7168,0.03926400012440152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,3584,0.316319121254815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,128,0.0029057777590221832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,128,0.0026622222115596137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,128,0.008776889079146916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,1024,0.006504888749784893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,256,0.010207110808955299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,8192,0.01181422256761127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,64,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,1536,0.10653155379825169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,7168,0.04752444558673435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,512,0.019914666811625164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,16384,0.011176000038782755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,64,0.0053084443012873335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,4096,0.007271111011505127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,2560,0.07410399781333075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,128,0.011494222614500256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,4096,0.026085333691702947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,3584,0.019105777144432068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,5120,0.009214222431182861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,7168,0.010467555787828235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,128,0.002708444371819496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,10240,0.13538489076826307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,256,0.013597333596812354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,12288,0.037676443656285606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,7168,0.15551911460028753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,65536,0.11640177832709418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,256,0.12346755133734809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,1536,0.005701333284378052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,1024,0.012436444560686747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,512,0.004554666578769684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,6144,0.005922666854328579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,2560,0.007132444116804335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,65536,3.3030461205376516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,65536,0.06829688946406047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,3072,0.07528355386522081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,5120,0.04760622315936618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,64,0.015641777051819693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,2560,0.007700444095664554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,3072,0.045371555619769625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,1024,0.00701155596309238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,128,0.004108444270160464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,128,0.002805333377586471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,128,0.003809777812825309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,2048,0.005960888746711943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,16384,0.024999999337726172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,768,0.0074746666683091064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,3584,0.01733688844574822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,32,0.005201777650250329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,1024,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,12288,0.021242666575643752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,2560,0.012784000072214337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,2048,0.008934222161769867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,128,0.005985777825117111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,2048,0.010199111368921068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,65536,0.12527467144860163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,4096,0.011134222149848938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,12288,0.01517777807182736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,32,0.05568799707624647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,4096,0.15966578324635824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,768,0.007734222544564142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,64,0.005503111001518037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,2560,0.005668444352017508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,768,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,3584,0.013961777091026306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,8192,0.03733155462476943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,2048,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,3072,0.007159999675220913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,512,0.008779555559158325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,2560,0.014589332871966891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,1536,0.012447111308574677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,1536,0.031585776143603854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,2048,0.011819555527634092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,64,0.008167111211352879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,7168,0.6966648631625705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,128,0.006451555424266391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,3072,0.018750222192870248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,1536,0.08628088898128933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,1024,0.004979555391603046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,2048,0.0058053334554036455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,5120,0.06328977478875054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,2560,0.023427555958429973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,12288,0.10990489191479152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,16384,0.06623288657930163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,1536,0.009176000124878353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,16384,0.02772888872358534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,1536,0.01185955521133211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,7168,0.008310221963458592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,12288,0.017867555220921833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,10240,0.010468444062603844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,256,0.004405333350102107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,768,0.0033946666452619764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,12288,0.0561751127243042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,256,0.005070222334729301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,1536,0.005199999858935674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,6144,0.014283556077215405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,10240,0.03895555602179633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,3584,0.00867377801073922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,256,0.004211555752489301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,1024,0.005882666756709416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,16384,0.026678222748968337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,16384,0.0730195575290256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,512,0.006588444527652528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,16384,0.0064080001579390625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,128,0.004983110974232356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,5120,0.02018666598531935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,256,0.006992888947327931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,2048,0.1788017749786377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,1536,0.008630221916569604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,768,0.008457777400811514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,5120,0.6982995669047037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,768,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,2560,0.0458222230275472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,3072,0.009331555830107795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,4096,0.011388444238238864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,256,0.002983999955985281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,6144,0.02001066671477424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,3584,0.007682667010360294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,512,0.005058666484223472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,2048,0.007347555624114142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,768,0.0076044441925154785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,12288,0.03831911087036133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,7168,0.028059555424584284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,3584,0.016746666696336534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,65536,6.506050533718533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,3072,0.01219466659757826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,128,0.007862221863534715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,1024,0.018231999542978074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,16384,0.0388008885913425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,65536,0.08839822477764553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,12288,0.05952799982494778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,65536,0.0770160026020474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,256,0.013361777696344586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,65536,0.25364798969692653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,3584,0.03641866644223531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,768,0.00600622221827507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,256,0.023690667417314317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,768,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,32,0.010614222122563256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,256,0.005069333232111401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,2048,0.006738666858938005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,256,0.005519999812046687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,3072,0.0075706665714581805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,12288,0.43322311507330996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,768,0.009379555781682333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,16384,0.023377777801619634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,10240,0.2797448900010851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,4096,0.01941155559486813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,12288,0.03250577714708116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,12288,9.903399997287327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,3584,0.013643555343151093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,32,0.005330666485759947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,64,0.006286222073766921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,5120,0.09060533179177178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,6144,0.016798221402698092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,6144,1.4471289316813152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,1024,0.008103110724025303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,64,0.007091555330488417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,3584,0.01163644426398807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,6144,0.015058666467666626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,256,0.007780444290902879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,3072,0.013820444544156393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,512,0.006205333189831839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,3584,0.02464177707831065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,5120,0.029271112547980413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,6144,0.03008355696996053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,8192,0.10481422477298313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,256,0.002945777856641345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,1536,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,2048,0.012921777864297232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,768,0.020656888683636982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,3584,0.028355555401908025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,6144,0.018763555420769584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,12288,0.02491644521554311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,8192,0.016559999850061204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,768,0.021086222595638696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,256,0.0029102222373088202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,4096,0.021913778450753953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,4096,0.008362666600280339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,12288,0.0304097765021854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,6144,0.014214222629865011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,2048,0.010608889162540436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,2048,0.059262222713894315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,65536,0.14029688305324978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,32,0.011071999867757162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,768,0.004794666750563515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,12288,0.061905774805280894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,2048,0.01610933409796821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,4096,0.0058702222175068325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,768,0.005397333453098933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,2560,0.0058782220714622075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,6144,0.01590133375591702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,4096,0.010751999914646149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,5120,0.012617777619096967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,65536,0.011052444577217102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,1536,0.011539555258221097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,5120,0.02310400042268965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,4096,0.013501333693663279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,7168,0.006270222365856171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,3072,0.09762933519151475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,1024,0.04930044545067681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,1536,0.1762817833158705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,7168,0.01313066648112403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,5120,0.04038755430115594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,7168,0.2714097764756945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,2560,0.04879466692606608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,256,0.004092444562249714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,1536,0.009663110805882348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,1024,0.014121777481502958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,128,0.01367377738157908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,10240,0.018933332628673978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,7168,0.012338666452301873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,128,0.0047333331571684945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,4096,0.008195555872387355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,1024,0.006000888844331105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,1536,0.0064239998658498125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,12288,0.04510666595564949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,5120,0.014058666096793281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,128,0.0027955555253558685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,3072,0.007616889145639207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,12288,0.376693328221639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,8192,0.030840890275107488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,2560,0.00815555536084705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,5120,0.02215555475817786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,1024,0.008485332959228093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,3584,0.011567999919255575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,8192,0.6264933480156792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,12288,0.018567999203999836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,128,0.004956444518433677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,1536,0.04634755518701342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,8192,0.011035555766688453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,6144,0.010135110881593492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,2048,0.010869333313571082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,2560,0.09245866537094116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,7168,0.00834222220712238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,1024,0.006502222269773483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,12288,0.054754667811923556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,3584,0.12368978394402398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,256,0.0036319999231232535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,128,0.0028479999552170434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,256,0.0033377777371141645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,65536,0.19130044513278535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,64,0.0058959999846087555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,8192,0.017602667212486267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,3584,0.016160888804329764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,3072,0.6905004183451334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,3584,0.007377777662542131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,768,0.0058293334311909145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,10240,0.058832890457577176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,12288,0.048057778014077075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,1536,0.006583111153708563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,3584,0.016520000166363187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,3072,0.01202755586968528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,10240,0.010007999837398529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,8192,0.00647199981742435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,2048,0.041504889726638794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,128,0.0027102221631341507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,1024,0.010203555226325989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,2560,0.00794400026400884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,64,0.0026462222966882917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,768,0.006252444452709622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,12288,0.020406223005718656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,5120,0.021381333470344543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,2048,0.051411555873023145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,768,0.0035235554807715942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,2560,0.011060444845093621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,128,0.0034293333689371743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,65536,0.01295644458797243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,7168,0.03474222289191352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,2048,0.006673777682913675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,1024,0.007186666958861881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,2560,0.011769777370823754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,65536,2.3984932369656033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,1536,0.02644088864326477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,12288,0.018325333793958027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,3072,0.10990666680865818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,1536,0.059281779660118945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,16384,0.06048266755210029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,64,0.008667555948098501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,2560,0.00905511114332411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,12288,0.8463315433926053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,12288,0.11159643861982559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,32,0.003524444583389494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,3584,0.021753778060277302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,4096,0.033230221933788724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,4096,0.10760977533128525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,2048,0.023355555203225877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,10240,0.00922400007645289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,6144,0.021800888909233943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,16384,0.023046221998002794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,5120,0.022499556342760723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,2560,0.012067555553383298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,64,0.06325422392951117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,10240,0.062344888846079506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,7168,0.02275377843115065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,1024,0.008512000242869059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,65536,0.12303466267055935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,2560,0.006057777752478917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,4096,0.011639110743999481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,12288,0.013263111313184103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,5120,0.0063697778516345555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,6144,0.022090666823916968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,3072,0.015580443872345818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,2048,0.005935110979610019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,12288,0.019517333971129525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,32,0.008758222063382467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,32,0.003907555507289039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,12288,0.015974221958054435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,5120,0.011622221933470832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,2048,0.0162435554795795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,64,0.004594666676388847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,8192,0.5515733295016819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,4096,0.01676177812947167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,32,0.005986666513813867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,32,0.0025920000755124623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,1536,0.0083137775460879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,12288,0.00998755544424057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,32,0.004790222065316306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,256,0.004239999999602635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,7168,0.023639111055268183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,512,0.0048071112897661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,3584,0.017654221918847825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,64,0.005491555564933353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,65536,0.07035466697480944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,2560,0.046366221374935575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,1536,0.006148444281684027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,3584,0.007871111234029135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,64,0.006629333313968446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,7168,0.06224089198642307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,3584,0.011016888750924004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,128,0.0041991112132867174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,16384,0.10685244533750747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,5120,0.013809777796268463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,512,0.005449777675999536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,10240,0.04572000106175741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,7168,0.011546666423479715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,128,0.0046968890560997855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,10240,0.2727546691894531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,16384,0.5962088902791342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,3584,0.007763555480374231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,7168,0.014542222023010254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,65536,0.27544355392456055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,8192,0.030221333106358845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,1024,0.01109333336353302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,65536,0.07775644461313884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,5120,0.012528888881206512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,2560,0.01204177737236023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,3584,0.017478222648302715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,4096,0.009992000129487779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,6144,0.25200533866882324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,64,0.06899822420544095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,64,0.002569777684079276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,12288,0.04273955689536201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,3072,0.008779555559158325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,1536,0.07307822174496122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,32,0.005159111072619756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,16384,0.03496888942188687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,64,0.0033982222278912864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,1024,0.006331555545330048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,65536,0.5395199987623427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,2048,0.016737777325842116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,10240,0.016858667135238647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,2048,0.005412444472312927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,4096,0.007770666645632849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,256,0.005140444470776452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,8192,0.020398222737842135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,256,0.0028880000528362063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,10240,0.06304977999793158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,4096,0.05633599890602959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,256,0.006170666466156642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,7168,0.020336889558368258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,5120,0.018333334061834548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,2560,0.007812444534566667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,512,0.003400000019205941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,8192,0.15644178125593397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,512,0.0047937776479456155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,768,0.006298666612969504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,3584,0.01822933389080895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,7168,0.014149333039919535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,3584,0.009696889254781935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,2560,0.10919911331600612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,512,0.006495999793211619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,65536,0.08023911052280001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,6144,0.0269768883784612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,2048,0.030270222160551283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,1536,0.015415110521846347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,7168,0.2789786656697591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,32,0.0033662221911880705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,3584,0.013697778185208639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,512,0.003195555466744635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,768,0.009962666365835402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,1536,0.005819555371999741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,1536,0.006664000037643645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,7168,0.02793955471780565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,2560,0.014763555592960782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,512,0.0052977779673205475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,3072,0.00959822204377916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,3072,0.030196444855795965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,16384,0.02334844403796726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,256,0.00272088891102208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,16384,0.05920088953442044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,256,0.00629333323902554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,7168,0.01185066666867998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,10240,0.0162124451663759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,65536,0.07137066788143583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,2560,0.0071937781241205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,256,0.004360888981156879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,2560,0.013472889032628802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,1536,0.009751111268997192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,10240,0.06468711296717326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,8192,0.029189334975348577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,1536,0.007296889192528195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,2048,0.00591111100382275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,512,0.016970665918456185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,65536,0.18339999516805014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,128,0.00407288885778851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,5120,0.011387555963463254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,4096,0.01205599970287747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,1536,0.005295111073387994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,5120,0.0793271130985684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,32,0.005402666827042897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,1024,0.013795555465751223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,12288,0.09999022218916152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,256,0.003491555651028951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,12288,0.03307111064592997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,2560,0.010527110762066312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,3072,0.016897777716318767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,128,0.004869333157936732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,256,0.010383110907342698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,256,0.004358222087224324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,2560,0.011855111353927188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,1024,0.0040942223535643685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,1024,0.012864000267452665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,65536,0.051664888858795166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,4096,0.08377599716186523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,1024,0.0052915554907586836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,2048,0.006952889263629913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,8192,0.017656889226701524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,32,0.007019555403126611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,128,0.0028524444335036804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,7168,0.06600177950329252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,512,0.006983111302057902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,128,0.004049777984619141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,4096,0.08070133129755656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,1024,0.007347555624114142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,16384,0.01844888925552368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,3072,0.021726222501860723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,1024,0.004148444367779626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,12288,0.049356444014443286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,32,0.024800888366169397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,10240,0.016540444559521146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,4096,0.009671111073758867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,10240,0.4611537721421983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,128,0.060644441180759005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,512,0.07589244180255465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,4096,0.022932445009549458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,65536,0.015728889240158927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,2048,0.021107556091414556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,6144,0.02345866627163357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,3072,0.05322666631804573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,12288,0.036144889063305326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,10240,0.0298977792263031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,12288,0.09890400038825141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,4096,0.028536889288160536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,2048,0.16120266914367676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,512,0.004283555679851108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,16384,0.02071466710832384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,3584,0.055123554335700146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,2048,0.006311999840868845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,2560,0.008526222573386298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,16384,0.01834488908449809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,8192,0.07121333148744372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,2560,0.006558222075303395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,1024,0.004924444274769889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,512,0.011922666596041786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,128,0.00518222235971027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,128,0.005858666780922148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,768,0.004831999954250124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,512,0.0036595554815398324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,8192,7.381966484917535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,2560,0.024658666716681585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,256,0.007134222322040134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,3072,0.011933333343929715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,512,0.010250666903124915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,32,0.002487999904486868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,5120,0.023248000277413264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,32,0.00452533322903845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,2048,0.13900355497996011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,8192,0.01651733285850949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,32,0.021551110678248938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,512,0.005415110952324337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,128,0.002587555597225825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,3584,0.01741066740618812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,1536,0.01458666721979777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,3072,0.10188711351818508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,128,0.005060444275538127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,3584,0.02843377656406826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,8192,0.40951021512349445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,256,0.006299555715587404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,12288,0.04313155677583483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,16384,0.020422221885787118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,32,0.04760533240106371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,32,0.0064533332155810455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,7168,0.048545777797698975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,3584,0.006855999843941794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,512,0.009653333160612319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,8192,0.012606222596433429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,3072,0.009372444616423713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,1024,0.011119999819331698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,256,0.009091555244392818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,4096,0.009583111438486312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,8192,0.029275556405385334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,10240,0.10534489154815674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,1024,0.04990399877230326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,512,0.0034808889031410217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,5120,0.009528888596428765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,6144,0.12034400304158528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,32,0.005335999859703912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,1024,0.005314666777849197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,256,0.0029724445194005966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,128,0.005203555441564984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,256,0.01089777797460556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,128,0.0047155556579430895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,512,0.005650666852792104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,8192,0.023426666855812073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,3584,0.06341777907477485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,32,0.004547555413511065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,7168,0.06322666671541002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,1024,0.005976000179847081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,128,0.00527022240890397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,128,0.004295999805132548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,16384,0.17752710978190103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,512,0.005442666510740916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,2560,0.007977777885066139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,65536,2.015380435519748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,1536,0.012969777815871768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,3072,0.0159262220064799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,256,0.004288889053795072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,2560,0.04121155540148417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,32,0.005099555684460534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,512,0.011312888728247749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,1536,0.015639111399650574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,16384,0.22881065474616158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,6144,0.6441609064737955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,64,0.003026666740576426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,8192,0.010015111002657149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,16384,0.02573866645495097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,1536,0.014035556051466199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,65536,0.07769244247012667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,3584,0.017052445146772597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,3072,0.016322667400042217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,7168,0.23617421256171334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,128,0.008369777765538957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,8192,0.03278844555219015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,2560,0.007710222568776872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,3072,0.008365333080291748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,3584,0.006863111009200414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,12288,0.02753066685464647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,512,0.005601777798599667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,512,0.026307556364271376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,1024,0.00851733320289188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,1024,0.004295999805132548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,512,0.004903111192915174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,3584,0.025560888979170058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,10240,0.02017688916789161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,256,0.0031902222997612427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,5120,0.009475555684831407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,12288,0.5921662118699815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,128,0.004036444342798657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,12288,0.04248799880345663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,32,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,128,0.0042151109211974675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,1024,0.08425955639945136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,12288,0.3683057890997992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,10240,0.018637332651350234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,6144,0.01586577792962392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,16384,0.028042667441897925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,1024,0.02600266701645321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,4096,0.1401040024227566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,10240,0.015305777390797934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,8192,0.026328888204362657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,8192,0.014294221997261047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,2048,0.19610310925377739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,64,0.006047111004590988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,16384,0.0720835526784261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,1024,0.004491555607981152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,3584,0.007795555724038019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,1536,0.006801777829726537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,8192,0.019367999500698514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,4096,0.009352000223265754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,6144,0.030461334519916113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,768,0.006697777658700943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,256,0.05152977837456597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,1536,0.005688000056478713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,512,0.016766221986876596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,2048,0.0872382190492418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,7168,0.016286222471131217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,5120,0.3361075454288059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,10240,0.028304000695546467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,65536,0.1996204455693563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,1536,0.045441776514053345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,5120,0.018904000520706177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,5120,0.0184560004207823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,8192,0.06582488616307576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,2048,0.0061919999619325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,2560,0.00833155545923445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,512,0.007567110988828871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,65536,0.3466035525004069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,128,0.004359999878538979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,12288,0.02327466673321194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,2048,0.006711999989218182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,5120,0.011036444041464063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,512,0.00869511150651508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,1024,0.003957333250178231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,64,0.003061333257291052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,32,0.0029519999192820657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,256,0.006243555496136348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,16384,1.1559315787421334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,65536,0.1485608948601617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,1536,0.006492444624503453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,10240,0.22272356351216635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,6144,0.032990223831600614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,64,0.0028248888750871024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,512,0.008674666285514832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,5120,0.8883395724826388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,1024,0.011615999870830111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,3584,0.009210666848553551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,5120,0.015023999743991427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,256,0.004675555560323927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,65536,1.234590212504069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,3072,0.00943911075592041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,3072,0.012460444536474017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,1024,0.0053048887186580235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,64,0.003072000005178981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,2048,0.022110222114457026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,12288,0.4207155439588759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,1024,0.01762488815519545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,3072,0.018041777941915724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,32,0.002571555475393931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,512,0.016709332664807636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,8192,0.016192888220151264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,12288,0.10765688949161106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,768,0.3093795511457655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,6144,0.9745306438869901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,32,0.002875555513633622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,10240,0.023271110322740342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,1536,0.004884444591071871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,65536,0.07470311058892144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,3072,0.006612444503439798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,65536,0.18395555019378662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,1024,0.007548444800906711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,10240,0.13472889529334173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,5120,0.010936000280910067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,3584,0.027657777070999146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,6144,0.02018044392267863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,7168,2.8561066521538625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,768,0.004690666579537922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,4096,0.012121777567598553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,768,0.007784888976150089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,12288,0.019884443945354886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,1536,0.03592355714903937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,1024,0.0057795556883017225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,7168,0.22856267293294272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,2048,0.0074035558435651995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,7168,0.44310487641228574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,1024,0.005013333426581489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,3584,0.11700533496008979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,2560,0.09700355264875625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,3072,0.009162666896979014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,7168,0.021147555775112573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,3584,0.3208071125878228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,2560,0.01241422196229299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,256,0.004471999903519948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,16384,0.07771378093295626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,7168,0.011500444677140979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,12288,0.022011554903454248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,1024,0.008948444492287105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,8192,0.015633778439627755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,65536,0.24147465493943956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,1024,0.0071937781241205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,5120,0.013321778012646569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,16384,0.05353155401017931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,64,0.002507555608948072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,256,0.002756444530354606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,65536,0.016755556066830952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,7168,0.03705777724583944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,1536,0.0052782222628593445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,512,0.0035911111368073356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,3584,0.01483022173245748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,65536,0.34580445289611816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,1024,0.008072888685597314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,256,0.007603555917739868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,10240,0.03527644607755873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,7168,0.008220444122950235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,1024,0.02604088849491543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,7168,0.017489777670966256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,3072,0.007072000039948358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,2560,0.009729777773221334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,256,0.005548444473081165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,2048,0.019344889455371432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,5120,0.012619555824332766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,10240,0.06417866547902425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,1024,0.010183999935785929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,10240,0.016764443781640794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,12288,0.017665778597195942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,768,0.006445333361625671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,65536,0.9898026784261068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,256,0.015817777978049386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,65536,0.09902933571073745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,12288,0.016286222471131217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,1024,0.007428444094128079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,65536,0.09512888722949558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,6144,0.028385778268178303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,512,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,1536,0.00867377801073922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,6144,0.026445334156354267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,12288,0.030184000730514526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,64,0.002941333378354708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,5120,0.009593778186374241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,16384,0.024157333705160353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,2560,0.0119991112086508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,256,0.004231111043029361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,128,0.011394666300879585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,128,0.0041848888827694785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,5120,0.31885867648654515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,10240,0.015928000211715698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,12288,0.0669377777311537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,5120,0.03446311089727614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,3072,0.00739555557568868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,4096,0.012503110700183444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,512,0.00435644429590967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,4096,0.031236443254682753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,10240,0.015760888655980427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,768,0.0047457776963710785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,1536,0.007100444700982835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,512,0.00850399997499254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,3584,0.01350488927629259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,32,0.009697777529557547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,16384,0.05115466647677951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,10240,0.013184888495339287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,8192,0.07868622408972846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,65536,0.12802044550577799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,768,0.010646222366227044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,12288,0.20360533396402994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,65536,0.06709777646594577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,32,0.006278222219811545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,1024,0.004005333201752769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,3584,0.010618666807810465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,7168,0.013456000222100152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,5120,0.04267200165324741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,10240,0.17198756006028917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,8192,0.015269334117571512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,3584,0.013586666848924426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,256,0.011702222128709158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,7168,0.13494843906826443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,6144,0.009656888743241629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,256,0.007156444092591603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,512,0.013248888982666863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,128,0.01405066748460134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,5120,0.1129431062274509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,16384,0.01975822283162011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,4096,0.18005155192481148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,16384,0.19863111442989775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,2048,0.006491555521885554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,10240,0.024878222081396315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,2560,0.015406222807036506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,512,0.0052622221410274506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,65536,0.08140977885988024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,128,0.005312888986534542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,768,0.008756444685988957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,12288,0.07774755689832899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,7168,0.010568000376224518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,8192,0.02327466673321194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,16384,0.05519733495182461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,5120,0.01643022232585483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,512,0.011747555600272285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,4096,0.019861333900027804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,1024,0.01200711147652732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,512,0.019113777412308585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,3072,0.007840000092983246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,4096,0.1316924492518107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,512,0.005022222383154763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,1024,0.006955555743641323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,8192,0.030037333567937214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,1536,0.012355555262830524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,64,0.004401777767472797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,128,0.005056000004212062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,10240,0.09417955742941962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,12288,0.09742399719026353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,65536,0.34769243664211696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,12288,0.10305688778559367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,3584,0.12279378043280707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,6144,0.02237777743074629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,6144,0.11909243795606826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,16384,0.022258667482270136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,12288,0.04109777675734626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,768,0.010729778144094678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,1536,0.0068408888247278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,6144,0.025395555628670588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,2048,0.010505778094132742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,7168,0.12439555592007107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,768,0.024021334118313257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,3072,0.2983395523495144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,1536,0.0053119998839166425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,2560,0.05611199802822537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,256,0.00315022220214208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,512,0.005818666683302985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,16384,0.08641955587599014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,7168,0.017464000317785475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,2048,0.1082106696234809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,3584,0.012437333663304647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,1024,0.008033778104517195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,3072,0.06991289059321086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,3072,0.00849955528974533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,65536,0.06653955247667101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,10240,0.026913776993751526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,6144,0.0129439996348487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,8192,0.03648622168434991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,16384,0.07764355341593425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,12288,0.031587554348839655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,7168,0.014001778430408902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,5120,0.028583112690183852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,1024,0.006310222049554189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,128,0.006211555666393704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,12288,0.09696977668338352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,256,0.0052933332820733385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,6144,0.009621333744790819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,8192,0.015126221709781222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,7168,0.1354853312174479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,1536,0.007902222375075022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,1024,0.02554399934079912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,128,0.01926488843229082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,5120,0.05554755528767904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,512,0.012031111452314587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,512,0.03451022174623277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,10240,0.03838488790724013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,256,0.004449777719047335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,4096,0.011381333072980246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,5120,0.011725333001878528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,65536,0.14899467097388372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,8192,0.020064888728989493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,8192,0.013910222384664746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,3072,0.011511999699804517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,8192,0.02892177634769016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,1536,0.011315555208259158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,64,0.002969777832428614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,3584,0.06199377775192261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,4096,0.006060444646411472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,65536,0.19000621636708578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,768,0.008960000342792934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,8192,0.0780622230635749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,64,0.0052675555149714155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,1536,0.034155554241604276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,2560,0.011427555647161273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,3072,0.005975111077229182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,4096,0.008279999925030602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,256,0.004492444296677907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,3072,0.009155555731720394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,1536,0.005708444449636672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,65536,0.5498204231262207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,1536,0.010770666930410596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,7168,0.029968000120586816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,10240,0.03432977861828274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,4096,0.19989421632554796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,1024,0.011815999945004782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,256,0.0030817778574095834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,512,0.024717332588301763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,5120,0.019430221782790292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,768,0.011232000258233813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,1536,0.010472888747851053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,2560,0.010991999672518836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,1536,0.020596444606781006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,2048,0.00794044468137953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,10240,0.022115555074479844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,3584,0.01824355622132619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,64,0.002749333365095986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,1024,0.018044445249769423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,8192,0.027494221925735474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,5120,0.01349955548842748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,8192,0.020516445239384968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,32,0.005678222411208683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,3584,0.11110933621724446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,128,0.0063573333124319715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,64,0.002855111120475663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,16384,0.1462844477759467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,16384,0.018280888597170513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,2048,0.006520888871616787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,3584,0.01129422254032559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,768,0.00452533322903845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,64,0.005467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,4096,0.04614933331807455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,64,0.004919111314747068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,12288,1.7554879718356664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,3072,0.01905777719285753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,65536,0.18851644463009307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,8192,0.027612444427278306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,1536,0.00852888905339771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,4096,0.012063110868136087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,4096,0.014523555835088095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,768,0.03512622250450982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,10240,0.007159999675220913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,128,0.003108444520168834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,5120,0.02008800043000115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,2560,0.007374222079912822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,256,0.0057502223385704895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,65536,0.09010399712456597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,1024,0.03583377930853102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,10240,0.9231893751356336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,10240,0.06658133533265856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,128,0.003585777762863371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,6144,0.014319999350441827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,4096,0.01610666679011451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,6144,0.052355554368760854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,4096,0.01641155613793267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,4096,0.04513511061668396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,1024,0.037124445041020714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,6144,0.21838755077785918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,2560,0.007096889118353526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,1024,0.036496000157462224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,65536,0.6689875390794543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,256,0.005236444373925527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,3072,0.0074942219588491656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,8192,0.011624889241324531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,12288,0.045553776952955455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,128,0.01088711122671763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,1536,0.005774222314357758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,2560,0.006478222293986215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,8192,0.014734221829308404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,1536,0.03560444381501939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,3072,0.013528889252079858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,8192,0.03172533379660712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,16384,0.022259554929203455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,8192,0.015229332778188916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,16384,1.6354355282253688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,1024,0.005063999858167437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,32,0.0032506665835777917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,2560,0.009751111268997192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,256,0.004930666751331753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,16384,0.05891288651360405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,16384,0.023892444041040208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,8192,0.054013331731160484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,6144,0.1887591150071886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,5120,0.03775555557674832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,7168,0.47272268931070965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,16384,0.055529776546690196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,2560,0.009520889156394536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,5120,0.012525333298577202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,1536,0.007754666937722101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,256,0.0036880001425743103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,2048,0.006505777852402792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,6144,1.5261564254760742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,3072,0.009884444375832876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,3584,0.0224906669722663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,2048,0.0052871112194326185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,7168,0.0562115576532152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,128,0.0047031111187405055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,128,0.014831110835075378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,512,0.005294222384691238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,65536,0.2843404346042209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,256,0.0041644444896115195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,5120,0.023246222072177466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,4096,0.009095999929640029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,8192,0.016371554798550077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,3584,0.007326222128338284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,16384,0.020543111695183646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,4096,0.06023110946019491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,128,0.0041200001206662925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,6144,0.024124445186720953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,3072,0.009762666291660732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,16384,0.08075822061962552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,12288,0.01887822151184082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,512,0.031509333186679415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,1536,0.008851555486520132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,256,0.0029964444951878656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,128,0.0027226667023367355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,6144,0.012264000044928657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,65536,0.09546489185757107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,1536,0.006962666908899943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,4096,0.00830844458606508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,128,0.005290666802061929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,10240,0.039706667264302574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,5120,0.03161599900987413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,3072,0.020239111449983384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,64,0.00332622230052948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,4096,0.0323022206624349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,10240,0.006153777655627992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,6144,0.03855289022127787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,8192,0.03345511025852627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,7168,0.9840052922566732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,2560,0.012763555679056378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,256,0.004498666773239772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,5120,0.011511999699804517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,256,0.010452444354693094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,32,0.0028035555862718155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,10240,0.006548444430033366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,4096,0.012005333271291522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,2560,0.007806222471925948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,2048,0.048551999860339694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,2560,0.0151555554734336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,12288,0.016866667403115165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,7168,0.23863199022081164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,256,0.01591199967596266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,256,0.1308915615081787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,768,0.0036657779581016968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,4096,0.07960710922876994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,65536,0.20480444696214464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,2560,0.1854382223553128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,8192,0.026234666506449383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,32,0.0026142222599850762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,12288,0.2226426601409912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,8192,0.052969776921802096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,2560,0.017403556240929496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,32,0.006350222147173352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,1024,0.0062977779242727495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,16384,0.04443111022313436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,8192,0.013882666826248169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,10240,0.021485333641370136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,2048,0.007661333514584436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,1024,0.012876444392734103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,1536,0.006865777903132969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,4096,0.009837333526876237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,128,0.0028284444577164124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,6144,0.019160888261265226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,2048,0.0070462218589252895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,16384,0.037195556693606906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,2560,0.005983110931184556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,2560,0.008319999608728621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,7168,0.07728888591130574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,5120,0.023544889357354905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,64,0.0025253333151340485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,256,0.00517155561182234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,16384,0.02338933282428318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,768,0.005671999934646819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,5120,0.011901333100265928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,5120,0.022416000564893086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,5120,0.036446223656336464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,8192,0.016879111528396606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,8192,0.015491555134455362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,512,0.009275555610656738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,3584,0.01422044469250573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,2048,0.011562666959232755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,5120,0.014312000738249885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,8192,0.01999288962946998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,7168,0.03420177764362759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,6144,0.013460444079505073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,768,0.030887110365761652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,768,0.031621333625581526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,1024,0.009187555147541894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,768,0.005106666849719154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,1536,0.00461422238085005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,4096,0.02557955516709222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,16384,0.019905777441130746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,2560,0.007439110842016008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,1536,0.08625155687332153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,4096,0.019476445184813607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,8192,0.019142222073343065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,1024,0.014879110786649915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,6144,0.025363556212849084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,256,0.004643555730581284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,32,0.002551111082235972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,256,0.004324444466167026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,2560,0.023543111152119104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,10240,0.01444355481200748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,8192,0.047981331745783486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,10240,0.035014222065607704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,2560,0.012257777982287936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,65536,0.3877413272857666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,10240,0.0181013329161538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,16384,0.15379644764794245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,128,0.005906666732496685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,512,0.005889777921968036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,64,0.003269333392381668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,6144,0.027448889282014634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,12288,0.06593510839674208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,10240,0.013350221845838757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,2048,0.007449777589903937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,1024,0.06444799900054932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,65536,0.11731021934085423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,128,0.005436444448100195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,512,0.004285333471165763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,2048,0.07192977931764391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,10240,0.6858711242675781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,2048,0.006331555545330048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,3584,0.008375999828179678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,2560,0.07741955253813002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,128,0.004676444249020683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,65536,0.2524435520172119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,3584,0.036152889331181846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,4096,0.00831822223133511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,2560,0.006777777853939269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,7168,0.011927111281288994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,12288,0.032808889945348106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,4096,0.07367466555701362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,10240,0.05635644329918755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,6144,0.02625333269437154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,128,0.00282577777074443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,10240,0.034918222162458636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,2048,0.007010666860474481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,768,0.004342222379313575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,5120,0.04918933245870802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,4096,0.06182222233878242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,128,0.0933235552575853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,3072,0.007567110988828871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,128,0.0042444442709287005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,2048,0.01841600073708428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,256,0.0032088889016045462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,2560,0.006004444426960415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,6144,0.017243555850452848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,3584,0.009509333305888707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,4096,0.011119999819331698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,65536,0.305424001481798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,6144,0.00997511131895913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,4096,0.0064479998416370815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,10240,0.7976586553785535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,2048,0.008006222546100616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,64,0.005740444279379315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,3072,0.013281777501106262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,4096,0.024122666981485155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,256,0.0036924444139003754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,10240,0.09609688652886285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,3584,0.008839111361238692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,3584,0.013586666848924426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,128,0.00434044458799892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,65536,0.5671902232699925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,6144,0.015059555570284525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,2560,0.006271111054552927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,6144,0.006065777606434292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,2560,0.007177777588367462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,16384,0.010217777556843227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,12288,0.17967822816636828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,10240,0.07818755838606092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,16384,0.08788266446855332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,128,0.00776088900036282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,1536,0.012887111140622033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,512,0.005684444473849402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,12288,0.011951111257076263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,128,0.013052444491121503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,2560,0.013841778039932251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,2048,0.0052888890107472735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,8192,0.019916445016860962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,2560,0.008343111309740279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,5120,0.11654844548967148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,12288,0.8189217779371473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,1536,0.07243377632564969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,1024,0.020769778225156996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,256,0.004320888883537716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,512,0.047232889466815524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,16384,0.02426133387618595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,1536,0.009560888840092553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,8192,0.045457777049806386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,64,0.3640542295244005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,2048,0.048948443598217435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,128,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,64,0.006992888947327931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,1536,0.020185778538386028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,32,0.006764444626039929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,16384,0.06313866376876831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,7168,0.013609777722093793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,7168,0.028875556257035997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,16384,0.0362337761455112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,3072,0.007840000092983246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,32,0.006552000012662675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,3072,0.00885511106914944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,12288,0.017238222890430026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,32,0.006698666761318843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,16384,0.45284530851576066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,128,0.003379555626047982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,1536,0.005685333162546158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,3072,0.006531555619504716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,32,0.003683555457327101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,65536,0.12223022513919407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,2048,0.011160000330872007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,256,0.008343111309740279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,7168,0.028880887561374243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,256,0.005323555734422471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,128,0.00442755553457472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,768,0.006159111029571957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,2560,0.009857777920034196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,64,0.0027155555370781156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,768,0.008488000267081791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,12288,0.8297680219014486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,1024,0.0057653333577844836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,128,0.0046897778908411665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,65536,0.07754222551981609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,1024,0.014371555712487964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,768,0.06905688842137654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,6144,0.013256889250543384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,12288,0.019909333851602342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,7168,0.04403733213742574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,16384,0.17347378200954863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,10240,0.02407111061943902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,8192,0.014564444621404013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,65536,0.07730133003658719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,16384,0.10561955637402004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,4096,0.013765333427323235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,6144,0.01650311052799225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,512,0.004864000197913912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,2560,0.008216000265545314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,6144,0.01332800007528729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,16384,0.02072888943884108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,10240,0.03798755672242906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,7168,0.027614222632514104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,10240,0.440046230951945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,256,0.0037519998020595978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,7168,0.03369422091378106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,16384,0.07532978057861328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,768,0.003579555700222651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,2048,0.02555022140343984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,2560,0.007028444773621029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,4096,0.05236000153753492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,8192,0.01735733283890618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,256,0.0032097777972618737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,512,0.015414221419228448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,2560,0.014053333136770459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,12288,0.05752978059980604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,4096,0.011495999991893768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,6144,0.027458666099442378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,3072,0.013311111264758639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,2560,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,128,0.004500444564554426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,512,0.29117600123087567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,65536,5.509223090277778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,16384,0.033618665403789945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,65536,0.34246667226155597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,768,0.01459022197458479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,10240,0.1484773291481866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,512,0.004920000003443824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,1024,0.05760800176196628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,64,0.005476444545719359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,768,0.005444444302055571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,3072,0.014363555444611443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,32,0.002605333303411802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,768,0.0038399998512532976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,6144,0.009948444035318162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,2560,0.006801777829726537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,8192,0.0524213347170088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,512,0.00360355567600992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,128,0.005159111072619756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,3072,0.016189333465364244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,512,0.009322666459613377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,65536,0.11087289121415879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,10240,0.06191466914282905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,4096,0.95595826043023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,8192,0.00982844415638182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,512,0.007978666987684038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,65536,0.016146666473812528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,1536,0.006699555450015598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,512,0.004761777818202972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,1024,0.007477333148320516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,768,0.005670222143332164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,5120,0.014884443746672736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,32,0.0026586666289303037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,768,0.0041804446114434134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,1536,0.006186666587988536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,10240,0.06849510802163018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,1024,0.006601777755551868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,12288,0.07093155384063721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,7168,0.013527111046844058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,256,0.003779555360476176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,64,0.0034462221794658234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,10240,0.010022222167915767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,64,0.005399111244413588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,128,0.004557333472702238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,3072,0.018184888694021437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,2560,0.030135999123255413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,768,0.00610133343272739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,10240,0.016154666741689045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,128,0.0027395555128653846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,32,0.004631999880075455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,6144,0.009742221898502773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,16384,0.045612444480260216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,16384,0.35341776741875547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,512,0.004595555365085602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,8192,0.01994044416480594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,12288,0.01814844376511044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,5120,0.013041777743233575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,2048,0.005398222141795688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,7168,0.14957066377003989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,256,0.007252444823582967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,12288,0.21801422701941597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,10240,0.017986666825082567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,10240,0.09056622452206081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,256,0.005109333329730564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,256,0.003176888864901331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,768,0.0038159998754660287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,512,0.003622222277853224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,2560,0.02548533346917894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,12288,0.212717334429423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,3584,0.010275555153687796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,768,0.003683555457327101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,6144,0.01261333293384976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,16384,0.039397332403394915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,256,0.008802666432327693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,32,0.01570311188697815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,1024,0.009521777431170145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,6144,0.02945599953333537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,65536,7.01726574367947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,16384,0.11216978232065837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,10240,0.01533777846230401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,1536,0.008745777938101027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,512,0.01017066670788659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,65536,0.07227289014392428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,3584,0.013820444544156393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,7168,0.014534221755133735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,7168,0.013367111484209696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,768,0.0040960001448790235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,32,0.040891554620530866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,16384,0.05600800116856893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,128,0.0027546667390399506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,3584,0.028569777806599934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,8192,0.024675556355052527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,512,0.005440000030729506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,10240,0.036766221125920616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,1536,0.00620888877246115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,2048,0.00674311113026407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,64,0.028090665737787884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,64,0.0027742222365405825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,768,0.005930666708283954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,32,0.0052560000783867305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,768,0.0035333335399627686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,10240,0.01907999979125129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,12288,0.08132977618111505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,128,0.005362666729423735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,32,0.006824888702895906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,3072,0.009748444788985783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,12288,0.02569777766863505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,4096,0.01759199963675605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,256,0.005442666510740916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,2560,0.036679112248950534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,4096,0.029382222228580054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,8192,0.01828266680240631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,16384,0.04159644577238295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,4096,0.008364444805516137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,256,0.012501333322789935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,64,0.002642666714058982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,6144,0.027052443888452318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,256,0.0348417792055342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,65536,0.06644444333182441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,5120,0.006100444330109491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,768,0.006060444646411472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,10240,0.04372888803482056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,8192,0.024614221519894067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,1536,0.052695112095938794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,768,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,6144,0.01849600010448032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,768,0.15016888247595891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,4096,0.019077334139082167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,64,0.011046222514576383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,512,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,16384,1.3130915959676106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,256,0.009813333551088968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,2560,0.007237333390447829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,3584,0.0074720001882976955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,4096,0.011362666885058085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,2560,0.012492444780137805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,16384,0.02369155486424764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,512,0.0034506666577524612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,128,0.006531555619504716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,5120,0.7095368703206381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,6144,0.01293955577744378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,6144,0.03216622273127238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,1024,0.006234666539563074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,3584,0.014225777652528552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,128,0.005084444251325396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,5120,0.10958488782246907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,128,0.0033679999825027254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,1024,0.009368889033794403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,32,0.003408000080121888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,16384,0.4360675546858046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,5120,0.01718311177359687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,32,0.005295111073387994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,3584,0.026649778087933857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,3584,0.04066933194796244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,1536,0.011764444410800934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,7168,0.010308444499969482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,65536,0.01290933373901579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,768,0.004851555658711327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,10240,0.01685422162214915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,4096,0.05597155623965793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,1024,0.006292444550328785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,2048,0.007098666495747036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,4096,0.06744800011316936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,1024,0.04925422204865349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,3072,0.026005332668622334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,3072,0.006697777658700943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,1024,0.012832000023788877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,128,0.006416000011894438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,3072,0.013614222407341003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,512,0.005597333527273602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,256,0.01758044461409251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,512,0.011857777833938599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,6144,0.012525333298577202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,6144,0.0064151109092765385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,2048,0.005315555466545953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,1024,0.028384000062942505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,2560,0.011864888999197217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,3584,0.015463999576038785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,768,0.006551110910044775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,3072,0.017264000243610807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,1024,0.005495111147562663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,768,0.009365333451165093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,16384,0.018001778258217704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,768,0.006214222146405115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,5120,0.011900444825490316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,4096,0.010507555471526252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,4096,0.07207022110621135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,32,0.003035555490189128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,10240,0.17547999487982857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,7168,0.006673777682913675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,65536,0.357752005259196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,64,0.028720888826582167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,12288,0.1161208947499593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,10240,0.01257066677014033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,64,0.009939555492666032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,256,0.004380444271696938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,7168,0.010885333021481832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,7168,0.013667555318938361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,16384,0.018671999375025432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,2048,0.060117330816056996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,2560,0.00904977818330129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,10240,0.10417955451541477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,6144,0.009887110855844285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,3584,0.007391999993059371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,6144,0.0609875586297777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,7168,0.10704622003767227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,128,0.004032888760169347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,256,0.003120888852410846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,4096,0.08238044712278578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,16384,0.06063377857208252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,6144,0.009684444301658207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,1024,0.006530666516886816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,512,0.004085333396991094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,4096,0.010320000350475311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,5120,0.006535111202134027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,5120,0.009897777603732215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,12288,0.04106488823890686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,128,0.024271999796231587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,12288,0.022862222459581163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,12288,0.01865333318710327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,512,0.0062871111763848206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,256,0.06866933239830865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,64,0.007831110722488826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,768,0.005696889013051987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,128,0.007065777977307637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,8192,0.01516711049609714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,4096,0.0417484442392985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,8192,0.03919644488228692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,768,0.004820444517665439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,10240,0.012624888784355588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,6144,0.013335111240545908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,7168,0.02775466607676612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,12288,0.021347555849287245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,3584,2.9712471432156033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,1024,0.007163555257850223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,3584,3.003935919867622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,7168,0.8504897753397623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,7168,0.01456622282663981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,2560,0.011876444849703046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,128,0.005416888743638992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,6144,0.015392000476519266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,64,0.008728000024954477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,10240,0.0646995570924547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,256,0.004713777866628435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,768,0.009224889179070791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,65536,0.06663111183378431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,32,0.0052897776994440295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,1024,0.004072000169091755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,2560,0.029653333955340918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,128,0.0046453335218959385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,6144,0.06667288806703356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,4096,0.011818666425016193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,1024,0.007994666695594788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,7168,0.15816799799601236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,2048,0.007059555914666917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,3072,0.006303999986913469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,128,0.004312889029582341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,2560,0.009465778039561378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,256,0.004777777940034866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,10240,0.017107556263605755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,10240,0.015754666593339708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,2560,0.10575022300084432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,6144,0.03490133417977227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,10240,0.3584684530893962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,7168,0.04530577858289083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,256,0.005107555538415909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,10240,0.01200888885392083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,128,0.0038906666967603895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,5120,0.00959644466638565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,512,0.015589333242840238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,2048,0.005472888963090048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,16384,0.03388177686267429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,4096,0.20186755392286512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,4096,0.022252443763944838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,256,0.002943111169669363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,16384,0.018547554810841877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,1536,0.0070462218589252895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,3584,0.007714666426181793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,2048,0.08110222551557753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,7168,0.02386666668785943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,6144,0.006510222123728857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,12288,0.031194667021433514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,6144,0.01496088835928175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,1024,0.006521777974234686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,2560,0.018963555494944256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,768,0.012773333324326409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,256,0.006492444624503453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,1536,0.010430222584141625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,128,0.005072000126043956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,3584,0.015138667490747241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,512,0.004513777792453766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,12288,0.0373288889726003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,5120,0.01684977776474423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,5120,0.033977776765823364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,4096,0.024875554773542616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,12288,0.02130222154988183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,16384,0.0329715543323093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,3584,0.16337154971228704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,65536,0.33226309882269967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,10240,0.01200355589389801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,16384,0.39098132981194395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,2048,0.008534222013420528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,768,0.035403556293911405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,1024,0.14793154928419325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,8192,0.03260533346070184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,7168,0.033244444264305964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,8192,0.029123554627100628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,32,0.0262097782558865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,768,0.006139555739031897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,3584,0.018036444981892902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,2560,0.006688000013430913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,6144,0.006361777583758037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,32,0.02130311065249973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,4096,0.017439999514155917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,768,0.004263999975389904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,8192,0.013422222601042854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,1024,0.010393777655230628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,512,0.007183999651008182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,8192,0.00683111117945777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,32,0.0027875554644399216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,128,0.007048889166778988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,16384,0.03135910961363051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,7168,0.04468533396720886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,8192,0.0574222207069397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,8192,0.029473778274324205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,256,0.0047884442740016514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,10240,0.017343110508388944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,768,0.008755555583371056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,3584,0.010963555839326648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,6144,0.0114702218108707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,3072,0.06171377499898275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,32,0.005373333477311664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,2048,0.01484444406297472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,16384,0.023364444573720295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,3584,0.017416889468828838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,65536,1.2587884267171223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,256,0.006725333217117522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,128,0.004426666845877965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,1536,0.006103111224042044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,256,0.004290666845109728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,8192,0.01587199999226464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,10240,0.07635733154084948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,2048,0.01874755488501655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,64,0.002583111118939188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,1024,0.003990222182538775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,16384,0.023620444867346022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,512,0.00545155546731419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,65536,4.859181298149957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,3072,0.014392889208263822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,512,0.05074933171272278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,4096,0.009383111364311641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,10240,3.0592577192518444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,8192,0.019810666640599568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,12288,0.029712888929578993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,12288,0.07811199956470065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,12288,0.6451350847880045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,1024,0.009299555586444007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,3072,0.0088755554623074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,3584,0.04562311040030586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,2560,0.022638221581776936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,7168,0.028909332222408716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,10240,0.01962844365172916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,12288,0.0229093333085378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,64,0.00406666679514779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,64,0.007840000092983246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,10240,0.023834667272037927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,1024,0.0052897776994440295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,12288,0.12477778063880073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,3072,0.007087111473083496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,6144,0.01899111105336083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,65536,0.06717333528730604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,2048,0.010428444378905825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,64,0.002809777855873108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,16384,0.8940693537394205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,7168,0.008110221889283922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,128,0.00517511119445165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,5120,0.028385778268178303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,1024,0.01145333300034205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,7168,0.2227706644270155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,7168,0.017646221650971305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,128,0.04432177874777052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,6144,0.0063164445261160536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,64,0.0029680000411139596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,2048,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,1024,0.01088444474670622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,2560,0.08057955238554212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,7168,0.015026667051845126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,5120,0.008671999805503422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,512,0.006239111224810283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,3072,0.007450666692521837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,768,0.005873777800136142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,3072,0.018744889232847426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,3584,0.029327111111746892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,2560,0.006148444281684027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,512,0.004009777886999978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,512,0.004838222016890843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,1536,0.00814666681819492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,10240,0.01864622202184465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,12288,0.03220977716975742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,1536,0.010199999643696679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,5120,0.017847999930381775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,12288,0.10263911220762466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,3584,0.01680088871055179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,5120,0.012758221891191272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,768,0.00371111101574368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,64,0.016594666573736403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,6144,0.05513244536187914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,65536,0.21207288901011148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,64,0.012892444100644855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,2048,0.0063777777055899305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,2048,0.009574222067991892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,65536,0.08092621962229411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,3072,0.011857777833938599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,4096,0.03321422139803568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,32,0.007095110913117726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,8192,0.01276088919904497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,10240,0.02120622164673275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,5120,0.012143111063374413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,12288,0.00618488879667388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,12288,0.1771306726667616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,4096,0.01032977799574534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,1024,0.0207360006040997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,64,0.003003555453485913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,4096,0.08025511105855306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,6144,0.07495466868082683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,5120,0.00869511150651508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,12288,0.01720622181892395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,128,0.0028239999794297745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,768,0.0075191110372543335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,3584,0.015799999237060547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,256,0.09943911102082993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,5120,0.01422044469250573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,10240,0.05269333389070299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,1024,0.010595555934641095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,7168,0.014329777823554145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,8192,0.05290844374232822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,3072,0.010026666853162978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,128,0.04804089003139072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,4096,0.09016088644663493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,3584,0.32507644759284127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,128,0.005164444446563721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,64,0.011128889189826118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,128,0.0030577776746617425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,1536,0.012525333298577202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,12288,0.019818666908476088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,256,0.039303110705481634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,1536,0.014210666219393412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,6144,0.06970933410856459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,6144,0.11606400542789036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,768,0.005873777800136142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,7168,0.02066044509410858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,10240,0.017866666118303936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,768,0.005718222094906702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,768,0.006755555669466655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,16384,0.04311555624008179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,12288,0.018744000130229525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,6144,0.02574933403068119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,3584,0.008036444584528605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,3584,0.007823111282454597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,5120,0.008797333472304875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,16384,0.021732444564501446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,7168,0.018952889574898612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,512,0.005923555543025334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,768,0.030039111773173015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,7168,0.12714044253031412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,1536,0.016363556186358135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,768,0.005620444400442972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,3584,0.00830488900343577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,8192,0.11211200555165608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,12288,0.015073777900801765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,2560,0.03203377789921231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,1024,0.0039048890272776284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,3584,0.006115555349323485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,8192,0.03035022152794732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,32,0.009611555271678502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,128,0.00281866660548581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,3584,0.00925511121749878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,7168,0.011720000041855706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,1024,0.030416889323128596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,128,0.002790222151411904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,256,0.01828799976242913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,1024,0.006031999985376994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,128,0.0027724444452259275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,6144,0.013248888982666863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,2560,0.009524444739023844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,1024,0.01421955558988783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,5120,0.012595555848545499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,5120,0.008615111311276754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,5120,0.008729777402347988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,3584,0.023335999912685815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,128,0.025215110845035974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,12288,0.14311022228664821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,768,0.00794311116139094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,4096,0.010216000179449717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,3072,0.042605333858066134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,12288,0.21187822024027506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,12288,0.027279110418425664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,1536,0.006463999963468975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,6144,0.01553066737122006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,16384,0.14009333981408015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,2048,0.006962666908899943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,768,0.004914666629499859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,3072,0.5759831004672581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,16384,0.08579377995596991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,5120,0.01695999999841054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,128,0.006744888921578725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,6144,0.02313066687848833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,3072,0.0092275556590822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,256,0.0028888889484935333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,65536,0.21648977862464058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,16384,0.5247839821709527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,8192,0.014869333969222175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,12288,0.07588177919387817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,8192,0.02530844509601593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,2048,0.08851999706692165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,1024,0.004006222304370669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,8192,0.7149164411756727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,10240,0.023676445086797077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,5120,0.008877333667543199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,512,0.02204088866710663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,1024,0.008972444468074376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,1536,0.005718222094906702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,1536,0.0057795556883017225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,5120,0.012876444392734103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,2048,0.006103111224042044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,768,0.0074311114019817775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,65536,0.9514302147759331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,7168,0.019447111421161227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,512,0.05337422092755636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,65536,0.06068533658981323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,32,0.005290666802061929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,10240,0.006228444476922353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,65536,0.014418666561444601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,2560,0.006879999819729064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,10240,0.01866311166021559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,3072,0.020033778415785897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,64,0.010038221875826517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,12288,0.17567733923594156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,2560,0.030304888884226482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,2048,0.014318222800890604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,5120,0.04695733388264974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,2560,0.006034666465388403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,7168,0.018188445104493033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,768,0.005046222358942032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,10240,0.014111111561457316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,128,0.0692968898349338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,128,0.005043555465009478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,128,0.006688888702127669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,12288,0.1464675532446967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,12288,0.015580443872345818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,8192,0.04236266679233975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,12288,0.060961776309543185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,3072,0.007551111280918121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,6144,0.020831111404630873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,1536,0.011579555769761404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,16384,0.011667555405033959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,7168,0.0810097787115309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,64,0.006601777755551868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,8192,0.01443911095460256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,12288,0.10234755277633667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,65536,0.4117422103881836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,2560,0.008247999681366814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,2048,0.006520888871616787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,8192,0.01535288823975457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,65536,1.5612817340426977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,2048,0.008794666992293464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,12288,0.007863110966152614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,5120,0.010412444670995077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,2560,0.01185155577129788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,16384,0.8228115505642362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,10240,0.3686248991224501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,10240,0.019102222389645047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,2560,0.00981955561372969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,128,0.007113777928882175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,1536,0.0047306666771570844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,32,0.002696000039577484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,1024,0.42918933762444395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,7168,0.34023022651672363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,256,0.01366488883892695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,3584,0.007930667036109501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,3584,0.08982755740483601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,16384,0.06118844615088569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,3072,0.02626400027010176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,2048,0.007035555938879649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,16384,0.01977244516213735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,65536,0.7604719797770182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,3584,0.12219377358754475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,6144,0.0270453327231937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,6144,0.10665600167380439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,3584,0.017634666628307767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,64,0.006808888994985157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,3072,0.01575999955336253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,768,0.0120319997270902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,32,0.005607999861240387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,1536,0.008244444098737504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,12288,0.030985779232449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,3072,0.008398222426573435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,3072,0.019870221614837646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,16384,0.3486044406890869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,768,0.01071733319097095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,2560,0.028360890017615423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,3584,0.00742222203148736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,5120,0.006647111227114995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,512,0.004978666702906291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,8192,0.025570667452282373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,2560,0.020991999242040846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,768,0.015067555838161044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,32,0.0026400000270869997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,16384,0.08054133256276448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,256,0.004960000101062986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,768,1.3200186623467338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,2560,0.06985866361194186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,2048,0.014519999424616495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,32,0.0031848889258172777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,256,0.009752888646390703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,7168,0.006412444429265128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,768,0.0037431112594074676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,6144,0.026299556096394856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,16384,0.030377778742048476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,2560,0.026498667067951624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,7168,0.1577155590057373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,128,0.005396444350481033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,12288,0.07765689161088732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,12288,0.006521777974234686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,3584,0.00793599999613232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,1536,0.014565333724021912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,256,0.004436444491147995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,1536,0.01624000072479248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,7168,0.11446577972835964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,3584,0.009280000295903947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,1536,0.004974222017659081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,5120,0.16771199968126085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,6144,1.1749519771999783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,1536,0.0063973334100511335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,5120,0.04369066821204292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,32,0.013251555462678274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,128,0.004491555607981152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,768,0.006188444379303191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,16384,0.12674310472276476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,3584,0.015959110524919298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,3584,0.03712977634535895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,1024,0.013199110825856527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,12288,20.044825236002605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,5120,0.029683556821611192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,128,0.004377777791685528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,1024,0.007096000015735626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,256,0.007575111256705389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,1536,0.006793777975771163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,4096,0.006535111202134027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,128,0.003914666672547658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,256,0.06590221987830268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,6144,0.02759733299414317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,5120,0.02539377742343479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,2048,0.016729777057965595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,64,0.0036968886852264404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,6144,0.02034133341577318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,256,0.004360888981156879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,256,0.01594666639963786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,1024,0.006001777946949005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,4096,0.00720355576939053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,32,0.005716444303592046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,2048,0.013191999660597907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,128,0.002999111182159848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,10240,0.45097780227661133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,1024,0.004119111018048393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,1024,0.01882577770286136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,64,0.007170666423108842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,128,0.03402311272091336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,3072,0.0058373332851462895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,10240,0.02743911080890232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,3072,0.00925688859489229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,1536,0.01846577723821004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,10240,0.011191111471917896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,512,0.005053333524200651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,2048,0.019345778557989333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,10240,0.013573333621025085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,6144,0.03515733281771342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,16384,0.271271997027927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,1536,0.23064888848198783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,3072,0.010556444525718689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,512,0.006477333191368315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,3584,0.01146577795346578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,8192,0.018923555811246235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,4096,0.045929776297675244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,32,0.0032764443506797156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,6144,0.014945778581831189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,3072,0.007838221887747446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,2560,0.0306915541489919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,2048,0.011800000237094032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,12288,0.029389331738154095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,2048,0.01181333346499337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,128,0.005381333331267039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,1536,0.00925155563486947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,16384,0.17739644315507677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,256,0.005147555636035072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,12288,0.006882666713661618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,768,0.003826666623353958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,64,0.027114666170544092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,8192,0.014871999621391296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,10240,0.028568890359666612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,7168,0.010412444670995077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,2048,0.018737778067588806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,6144,0.09199111329184638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,8192,0.031055112679799397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,5120,0.010695111420419482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,7168,0.00654666663871871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,16384,0.05987733602523804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,5120,0.011775111158688864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,7168,0.3666364351908366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,5120,0.0125591109196345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,3072,0.47285064061482746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,1024,0.005583111196756363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,32,0.006555555595291986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,7168,0.01683555543422699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,128,0.0041422223051389056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,64,0.004291555533806483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,32,0.0032551110618644287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,2048,0.012830221818553077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,65536,0.11025777790281509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,6144,0.017716444200939603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,6144,0.02474400069978502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,3584,0.008958222137557136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,2048,0.0059573331640826324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,3584,0.00978311151266098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,3072,0.010123555858929953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,3584,0.007292444507280986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,5120,0.08559111091825698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,1024,0.008418666819731394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,3584,0.024655999408827886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,2048,0.01593155496650272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,2560,0.015816888875431485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,7168,0.017507554756270517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,64,0.005409777578380372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,32,0.00610133343272739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,3072,0.02214844524860382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,2048,0.20867821905348036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,512,0.0034311111602518293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,3584,0.011029333704047732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,128,0.007646222081449296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,512,0.006467555546098285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,2048,0.007830222447713217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,5120,0.021010667085647583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,768,0.006513777706358168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,512,0.018170666363504197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,12288,0.059565332200792104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,2560,0.08349421951505874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,512,0.005085333353943295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,512,0.004560889055331548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,5120,0.02588533361752828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,3072,0.030427555243174236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,32,0.005190222213665645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,6144,0.13424000475141737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,7168,0.07016088565190633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,2560,0.006015111174848344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,256,0.0033333332588275275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,3072,0.04856977860132853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,16384,0.24078223440382215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,256,0.032504889700147845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,65536,0.05942844682269626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,3584,0.02716800073782603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,8192,0.03506133291456435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,3584,0.012628444366984896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,256,0.00433955548538102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,2048,0.01128355579243766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,7168,0.01482311056719886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,8192,0.007337777978844113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,7168,0.1581795612970988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,1536,0.0047004446387290955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,3072,0.006785777707894643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,128,0.004008888784382078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,4096,0.01885422236389584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,1536,0.016128000285890367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,7168,0.5342942343817817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,65536,0.5411244498358833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,6144,0.03366311060057746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,5120,0.010519111322032081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,10240,0.04602933261129591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,3584,0.009782222410043081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,65536,0.06933955351511638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,6144,0.2714960045284695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,7168,0.007081777685218387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,2048,0.011325333681371478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,16384,0.02299022177855174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,65536,0.561997307671441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,7168,0.01958666741847992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,16384,0.04043022129270766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,32,0.01980799933274587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,10240,0.27047822210523814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,512,0.00618488879667388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,16384,0.015139554937680563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,16384,0.020576889316240948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,4096,0.10446755753623115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,12288,0.07408266597323947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,4096,0.006695111178689533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,32,0.00544088871942626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,256,0.00813333359029558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,7168,0.026343110534879897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,65536,0.06617866622077094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,5120,0.006834666762087081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,65536,0.06682488653394911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,64,0.0027751111321979095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,2560,0.014271999398867289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,8192,0.03088444471359253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,7168,0.01219377749496036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,768,0.021458667185571458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,5120,0.10074400239520603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,1536,0.008838222258620793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,12288,0.018058665924602084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,128,0.004262222184075249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,256,0.00721244431204266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,4096,0.015498666299713982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,1536,0.011824000212881299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,12288,0.0566506650712755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,1024,0.09880799717373318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,768,0.008030222521887885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,2560,0.5142951011657715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,2560,0.016114667057991028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,5120,0.021551999780866835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,1024,0.006813333266311222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,7168,0.01089244418674045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,65536,0.4042666753133138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,1536,0.00655733338660664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,512,0.011955555942323474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,6144,0.059431109163496226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,512,0.003415999934077263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,6144,0.4849093225267198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,3072,0.009047110875447592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,2560,0.0087013335691558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,1536,0.020376000139448378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,2560,0.04284977912902832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,512,0.00480000012450748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,6144,0.00998933364947637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,3584,0.05227111114395989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,1024,0.006317333214812809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,2560,0.008184889124499427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,8192,0.08371466398239136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,512,0.003232888877391815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,10240,4.138045416937934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,3072,0.008566222257084316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,3584,0.5529520246717665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,512,0.005103999955786599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,5120,0.017877333694034152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,6144,0.02689244515366024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,65536,0.08315377765231662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,3072,0.22198222743140328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,2560,0.0232968893316057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,1536,0.011044444309340583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,65536,7.440364413791233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,768,0.007990222010347579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,65536,0.330351988474528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,128,0.011767110890812345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,8192,0.03652000096109178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,4096,0.009757333331637913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,1024,0.005678222411208683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,5120,0.010671999719407825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,16384,0.06145066685146756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,256,0.05541066659821404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,5120,0.008193777667151557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,16384,0.1228497823079427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,1024,0.005144888742102517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,768,0.005666666560702854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,2048,0.01070755554570092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,6144,0.010073777702119615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,64,0.0029271110478374693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,768,0.007941332956155142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,1024,0.0121742222044203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,3584,0.007154666715198093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,4096,0.3706631130642361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,2560,0.016922665966881644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,12288,0.005956444475385878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,10240,0.03595111105177138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,64,0.0027226667023367355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,8192,0.015569777952300178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,512,0.004416000097990036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,768,0.012025777664449481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,8192,0.031010667483011883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,64,0.0028275555620590844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,7168,0.12521778212653265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,1536,0.2939271132151286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,1536,0.017306667235162523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,4096,0.1538791126675076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,65536,0.23992088105943468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,8192,0.009071110851234859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,3584,0.019318222999572754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,32,0.006690666493442323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,6144,0.01053155544731352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,5120,0.010829333629873065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,128,0.00442755553457472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,6144,0.013361777696344586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,64,0.01086400035354826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,2560,0.006554666492674086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,3584,0.014861333701345654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,7168,0.2351262304517958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,768,0.003532444437344869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,768,0.005712888720962737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,7168,0.34108710289001465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,768,0.050364444653193154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,1024,0.014740443891949125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,512,0.018570666511853535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,4096,0.04602488875389099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,7168,0.0480915539794498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,6144,0.04078133238686456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,5120,0.01591111057334476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,5120,0.14215555455949572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,65536,0.14154844813876682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,5120,0.025519111090236243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,1024,0.01707555519209968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,4096,0.02052177819940779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,3072,0.009854222337404886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,512,0.008941333327028487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,3584,0.00923288861910502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,12288,0.5370186699761285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,5120,0.015398222539159985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,7168,0.01610933409796821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,7168,0.2364035447438558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,3584,0.07579822010464139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,65536,0.19017155965169272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,3584,0.007715555528799693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,1024,0.02387911081314087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,1024,0.011515555282433828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,12288,0.078402664926317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,3584,0.02664533257484436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,2048,0.009103111094898647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,16384,0.10581066873338486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,3072,0.009158222211731805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,256,0.004909333255555895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,65536,0.610464890797933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,8192,0.016650666793187458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,1024,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,12288,0.00846400029129452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,7168,0.011035555766688453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,4096,0.011163555913501315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,32,0.02409511142306858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,10240,0.024244444237815008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,5120,0.020920000142521326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,5120,0.024632889363500807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,2560,0.006245333287451003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,512,0.05606310897403293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,4096,0.007743111087216272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,256,0.004578666554556953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,16384,0.08271111382378472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,7168,0.012812444733248817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,3584,0.017278222574128043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,4096,0.005935110979610019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,7168,0.007161777880456712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,256,0.002921777880854077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,6144,0.02511466708448198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,3584,0.00980000032318963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,8192,0.019267555740144517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,10240,0.008452444440788692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,8192,0.015019555886586508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,128,0.0063164445261160536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,1024,0.040291554398006864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,5120,0.009214222431182861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,2560,0.09053599834442139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,3072,0.006783999916579988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,65536,0.07298933135138617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,12288,0.10998755693435669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,3072,0.006811555474996567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,3072,0.00703999979628457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,10240,0.023388443721665278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,8192,0.014302222265137566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,1536,0.00646666685740153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,256,0.005488889084921942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,768,0.005474666754404704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,7168,0.04152266515625848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,2048,0.009530666801664565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,5120,0.01161688897344801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,6144,0.01142133358452055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,256,0.0054551110499435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,128,0.0028808888875775864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,8192,0.031184001101387873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,5120,0.02531911101606157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,2048,0.012423999607563019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,3072,0.01665511065059238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,5120,0.12564888265397814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,32,0.0029004443850782183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,16384,0.006084444622198741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,10240,0.018050667312410142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,4096,0.01016444464524587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,12288,0.0403457780679067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,8192,0.02215822206603156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,4096,0.3641822338104248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,8192,0.016784888174798753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,2560,0.008447111480765873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,5120,0.014344889256689282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,7168,0.1226373381084866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,10240,0.07704977856742011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,5120,0.015807110402319167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,256,0.005379555539952383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,3072,0.010060444474220276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,5120,0.026504889130592346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,2048,0.012984000146389008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,2560,0.029871109459135268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,7168,0.014225777652528552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,256,0.004290666845109728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,512,0.006105777704053455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,128,0.00618577789929178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,16384,0.046182221836513944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,128,0.39479732513427734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,128,0.007846222155623967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,10240,0.04121422105365329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,2560,0.16690399911668566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,16384,0.026345777842733595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,32,0.002565333412753211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,64,0.0025920000755124623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,12288,0.006434666613737742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,512,0.004833777745564778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,1536,0.028852442900339764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,5120,0.015970667203267414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,6144,0.048735111951828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,768,0.00792266676823298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,16384,0.03315022256639268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,512,0.00574933323595259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,7168,0.012588444683286877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,3072,0.01956888867749108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,12288,0.04327555497487386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,3072,0.01832355558872223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,12288,0.005969777703285217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,256,0.003261333331465721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,2560,0.052314665582444936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,4096,0.020208888583713107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,6144,0.018473777506086562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,10240,0.015171556009186639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,512,0.005782222168313132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,2560,0.015170666906568738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,6144,0.006783999916579988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,3072,0.035735110441843666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,16384,0.10274310906728108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,16384,0.04974844389491611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,3584,0.016585777203241985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,512,0.008085333638721043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,512,0.0038568890757030914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,512,0.007793777518802219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,768,0.007450666692521837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,768,0.006396444307433234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,64,0.0032711111836963226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,128,0.0027599999060233435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,64,0.0026480000880029467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,512,0.007224889265166388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,512,0.005231110999981563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,1536,0.004642666627963384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,8192,0.036791112687852644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,512,0.007845333053006066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,7168,0.014827556080288358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,65536,0.5414266586303711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,3072,0.012451555165979596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,7168,0.015576000014940897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,16384,0.020558221472634208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,768,0.005892444401979446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,7168,0.014408888088332282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,4096,0.02558044426971012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,65536,0.07583644655015734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,2048,0.0053191110491752625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,3072,0.01090755561987559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,5120,0.007706666986147563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,10240,0.00620977787507905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,8192,0.006125333408514659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,7168,0.015174221661355762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,768,0.008001777860853408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,16384,0.018029333816634286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,512,0.005380444642570283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,128,0.004840888910823398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,7168,0.029716445340050593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,2560,0.013769778112570444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,6144,0.022486221459176805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,65536,0.15234843889872232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,768,0.0052737775776121355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,768,0.006073777874310811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,3584,0.024873778223991394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,65536,0.07040266858206855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,64,0.009747555686367882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,32,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,2048,0.006153777655627992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,7168,0.02016266683737437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,2560,0.006507555643717448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,512,0.011411555939250521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,8192,0.0602160029941135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,768,0.011380444798204633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,1024,0.05111911230617099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,10240,0.029248889949586656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,12288,0.03745155533154806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,768,0.028910222980711196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,2560,0.013219555219014486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,6144,0.014182221558358936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,64,0.0038382220599386427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,1536,0.0053839998112784485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,5120,0.029231998655531142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,128,0.003891555385457145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,4096,0.008759111166000366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,6144,0.022038223014937505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,65536,1.4364097383287218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,12288,0.017941332525677152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,1536,0.006492444624503453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,5120,0.010045333041085137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,6144,0.02358755634890662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,1024,0.004371555729044808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,3072,0.006401777681377199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,256,0.10066933102077907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,10240,0.015031110909250049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,5120,0.02345866627163357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,10240,0.0181377778450648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,256,0.028375112348132666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,10240,0.01682133310370975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,3584,0.00909511082702213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,8192,0.055403557088640004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,10240,0.21859378284878203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,3072,0.007560888926188151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,8192,0.01648000048266517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,512,0.004997333304749595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,3072,0.007000889215204451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,65536,0.1914986636903551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,32,0.003718222181002299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,65536,0.07388710975646973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,1536,0.008163555628723567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,6144,0.16798310809665254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,65536,0.07692977454927233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,7168,0.053499556250042386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,4096,0.0414613319767846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,1536,0.011550222006109027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,2048,0.018254222141371835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,768,0.004893333133724001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,3584,0.010767111347781287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,8192,0.0074577778577804565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,128,0.003912888881233004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,7168,0.024150222539901733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,65536,2.411181343926324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,7168,0.010457778142558204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,3072,0.01034222212102678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,16384,0.10732444127400716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,65536,0.06590933269924588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,2048,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,3072,0.1424648894204034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,256,0.00434400017062823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,256,0.004858666823969947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,8192,0.006600889066855113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,256,0.004573333180612988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,1024,0.009895111123720804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,2560,0.0061919999619325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,512,0.0033173333439562055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,64,0.005122666557629903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,32,0.005021333280536863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,12288,0.010049777726332346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,512,0.00906133320596483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,10240,0.009758222434255812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,12288,2.1863110860188804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,7168,1.3992604149712458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,3072,0.05520800087187025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,8192,1.2098169326782227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,512,0.007045333584149678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,10240,0.05302222238646614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,256,0.009316444396972656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,12288,0.06390577554702759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,256,0.003236444460021125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,10240,0.17605777581532797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,512,0.007080888582600488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,65536,0.06557510958777533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,1024,0.007144889069928064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,8192,0.03176711002985636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,768,0.005176888985766305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,6144,0.012200888660218982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,7168,0.028386665715111628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,8192,0.2542791101667616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,1536,0.0734035571416219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,16384,0.020217777954207528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,10240,0.01737866633468204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,6144,0.01074577785200543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,5120,0.030272887812720403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,65536,0.17515910996331108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,7168,0.021700445148679946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,2048,0.006117333140638139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,6144,0.023656888140572443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,1536,0.0047591109242704184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,6144,0.025772444076008264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,7168,0.04211911227968004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,3584,0.00758133331934611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,64,0.010443555812040964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,128,0.0037191112836201987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,128,0.004559999952713649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,64,0.002609777781698439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,16384,0.12316089206271702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,10240,0.015249777171346875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,3584,0.007140444384680853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,32,0.003169777699642711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,12288,0.019070222973823547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,4096,0.038318223423428006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,3584,0.2842551072438558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,7168,0.026644445127911035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,2048,0.006640888750553131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,32,0.004957333207130432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,1536,0.012991111311647626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,256,0.005691555639108022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,1024,0.006439111298984951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,10240,0.05443911088837517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,512,0.00997688869635264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,5120,0.010326222413116032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,3072,0.013387555877367655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,10240,0.24258844057718912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,5120,0.008236444658703273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,32,0.006288888967699475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,6144,0.01422133379512363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,2048,0.022823111878501043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,128,0.003089777711364958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,512,0.005140444470776452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,5120,0.006776000062624614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,6144,0.005923555543025334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,1024,0.008593777815500895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,768,0.005200888961553574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,12288,0.06989777750439115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,2048,0.3752853340572781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,32,0.005395555661784277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,6144,0.00684799998998642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,8192,0.02908710969818963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,12288,0.025194666451878015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,16384,0.030586666531032983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,16384,0.6356897883945042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,512,0.01202222208182017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,6144,0.009815110928482479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,256,0.0058408888677755995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,3584,0.00871377769443724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,6144,0.020063999626371596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,256,0.011564444336626263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,256,0.0063982220987478895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,5120,0.08790489037831624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,2048,0.018575999471876357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,65536,0.06707021925184461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,2560,0.008407110969225565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,768,0.0058453331391016645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,128,0.005256888767083486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,5120,0.021976888179779053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,128,0.0030666666312350165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,256,0.00535111129283905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,7168,3.0369919670952688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,256,0.005704000178310607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,7168,0.022064000368118286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,256,0.0030373332815037835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,1536,0.00996888925631841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,256,0.006014222072230444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,64,0.002838222309947014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,2560,0.006346666564544042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,3072,0.01995466649532318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,1536,0.006312888943486744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,7168,0.038943999343448214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,1024,0.04074755642149184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,128,0.00507377791735861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,3072,0.22594134012858072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,10240,0.014208889669842191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,64,0.00757155567407608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,5120,0.0744204454951816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,64,0.002675555646419525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,768,0.008409777449236976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,4096,0.013585777746306526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,768,0.016511110795868766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,8192,0.03119644522666931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,256,0.013325333595275879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,7168,0.01332355539004008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,768,0.004920000003443824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,64,0.002605333303411802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,3584,0.010671999719407825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,1024,0.006780444333950679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,3584,0.10545600122875637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,256,0.017628444565667045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,8192,0.007344889144102733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,1024,0.024490666058328416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,10240,0.01627288924323188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,64,0.0027582223216692605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,1024,0.03269777695337931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,16384,0.018704000446531508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,65536,0.1037751105096605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,6144,0.015574221809705099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,512,0.0057297779454125305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,2560,0.014362666341993542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,3584,1.3775946299235027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,65536,0.09903911087248062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,128,0.004655111167165968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,6144,0.04798666636149088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,2048,0.00832799987660514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,128,0.002962666667169995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,65536,0.020599111914634705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,65536,0.14171111583709717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,8192,0.043749332427978516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,10240,0.05292355683114794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,2560,0.010296888649463654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,64,0.0029440000653266907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,2560,0.013235555754767524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,10240,0.06613510847091675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,7168,0.0148053334818946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,512,0.006095110956165526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,10240,0.0498924454053243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,12288,0.04651733239491781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,512,0.005629333357016246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,128,0.004280888785918553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,6144,0.024269334144062463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,5120,0.01979822251531813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,32,0.0025617778301239014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,65536,0.48533068762885195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,1536,0.005151999907361137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,65536,0.07111910978953044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,4096,0.3623342249128554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,3072,0.008098666866620382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,65536,0.25201866361829967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,8192,0.03165510959095425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,7168,0.03273422188229031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,2560,0.009006222089131674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,12288,0.04219111137919956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,5120,0.02843466732237074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,65536,0.3679857783847385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,65536,3.2667840321858725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,768,0.013247110777431063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,5120,0.01667644414636824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,3584,0.0074728892909155945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,128,0.004022222426202563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,6144,0.013025778035322825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,4096,0.008644444247086843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,6144,0.1931840048895942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,12288,0.04221422142452664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,64,0.005292444593376583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,12288,0.018838221828142803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,6144,0.00655644428398874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,8192,0.07408889134724934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,3584,0.01734044485621982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,65536,0.23108177714877656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,512,0.0033448889023727844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,1536,0.005394666559166378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,32,0.006011555592219035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,256,0.00461777796347936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,8192,0.028669334120220605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,512,0.004968889057636261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,2560,0.01036622209681405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,65536,0.0684373312526279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,2560,0.014780445231331719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,3584,0.013280889226330651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,2560,0.012111110819710625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,65536,0.11707999971177842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,128,0.004975999808973736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,3072,1.199096891615126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,32,0.003400000019205941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,1536,0.008535111116038429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,10240,0.016421332955360413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,12288,0.006370666540331311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,65536,0.1418479945924547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,3584,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,1536,0.0047031111187405055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,16384,0.023286221755875483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,12288,0.04834489027659098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,65536,0.06640266709857516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,32,0.01143022212717268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,3072,0.01297155519326528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,10240,0.021077333225144282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,1536,0.00980799976322386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,4096,0.010606221854686737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,256,0.01423822177780999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,5120,0.217684441142612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,1536,0.007339555356237624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,4096,0.006340444501903322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,2560,0.11770578225453694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,12288,0.058006220393710665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,2048,0.006806222101052602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,3584,0.010012444522645738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,10240,0.006575999988449945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,10240,0.03935022155443827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,128,0.005241777747869492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,1024,0.009887999958462184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,128,0.009856888817416297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,16384,0.015255999233987598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,16384,0.0063822223908371395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,6144,0.012950222525331708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,1536,0.009702222214804756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,12288,0.0185199992524253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,16384,0.27595554457770455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,10240,0.01310400002532535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,8192,0.013112000293201871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,1024,0.010300444232092964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,5120,0.014051554931534661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,6144,0.028930667373869155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,8192,0.046241776810752026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,7168,0.007858666280905405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,16384,1.7711750666300456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,768,0.02272266646226247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,128,0.0032142222755485107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,3072,0.008978666530715095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,8192,0.014101333088344999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,8192,0.01424888935354021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,2048,0.014284445179833306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,65536,0.08867200215657552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,5120,0.05892088678148058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,64,0.010812444819344414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,512,0.004965333475006951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,1536,0.15731467141045466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,5120,0.0672337810198466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,512,0.005004444470008214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,4096,0.11600977844662136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,3072,0.006792000184456508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,64,0.010235555469989777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,128,0.0028684443483750024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,4096,0.015287111202875773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,10240,0.33992711702982586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,4096,0.018202667435010273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,128,0.005323555734422471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,1024,0.043474667602115206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,512,0.0059262220230367445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,5120,0.053844445281558566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,3584,0.0092622223827574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,3584,0.02498133314980401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,65536,0.24980266888936362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,5120,0.013447999954223633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,5120,0.01018488903840383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,4096,0.27252888679504395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,16384,0.01534577707449595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,12288,0.03137244449721442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,7168,0.09462755256228977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,3072,0.008237332933478886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,2560,0.016930666234758165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,128,0.005438222239414851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,10240,0.0413262214925554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,768,0.007361777954631382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,4096,0.19213422139485678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,10240,1.0105297300550673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,12288,2.5104427337646484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,4096,0.01844888925552368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,8192,0.012496000362767113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,12288,0.14335821734534368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,5120,0.024302222662501868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,3584,0.02160444524553087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,6144,0.01797777745458815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,6144,0.012681778106424542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,256,0.0046924443708525766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,512,0.003240888938307762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,3072,0.019202666150199044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,1536,0.00816888858874639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,2560,0.02777066661251916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,5120,0.04937422275543213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,2048,0.011383111278216044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,4096,0.013717333475748697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,512,0.005565333283609814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,1536,0.02176888949341244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,3584,0.011598221957683563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,8192,0.019055111540688407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,5120,0.09879910945892334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,64,0.005625777774386936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,10240,0.014503111441930136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,768,0.005884444548024072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,5120,0.029117332564459905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,65536,0.11085422171486749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,768,0.005897777775923411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,3072,0.09642399681939019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,3584,0.01689511040846507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,4096,0.008345777789751688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,12288,0.028026666906144884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,256,0.004618666652176115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,1024,0.0058328890138202244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,65536,0.39576265547010636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,1024,0.08998488717608982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,8192,0.038655112187067665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,128,0.00272088891102208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,7168,0.02540622154871623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,12288,0.15843644407060412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,16384,0.35262756877475315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,5120,0.019158222609096102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,10240,0.0381297767162323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,1536,0.0072266666425599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,128,0.013058666553762225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,65536,8.332090589735243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,16384,0.04339111182424757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,8192,0.024650666448805068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,32,0.01791555517249637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,4096,0.02541422181659275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,32,0.002875555513633622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,10240,0.22770044538709852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,12288,1.1829004287719727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,8192,0.01757866640885671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,5120,0.030380444394217596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,16384,0.492611567179362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,256,0.003167111012670729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,16384,0.2629351086086697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,2560,0.00759733302725686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,16384,0.06502044200897217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,2048,0.03417866759830051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,2560,0.014462222655614218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,1024,0.009695111049546136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,1536,0.014718222949239941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,12288,0.015209777487648858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,2048,0.011728888584507836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,10240,0.018188445104493033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,4096,0.09289955430560642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,128,0.0027013334135214486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,12288,0.006868444383144379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,1536,0.005607111172543631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,3584,0.007487110793590546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,8192,0.15226044919755724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,1536,0.004616888860861461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,256,0.004366222355100843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,512,0.00508088908261723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,8192,0.011785777906576792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,6144,0.015471999843915304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,16384,0.3579395612080892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,768,0.006430222342411677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,1536,0.1251573297712538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,5120,0.01677600045998891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,2048,0.009348444640636444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,768,0.005075555708673266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,2560,0.005959110955397288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,512,0.25161422623528373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,256,0.006304889089531369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,1536,0.011805333197116852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,1024,0.08987555238935684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,2048,0.013012444807423485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,8192,0.06568977567884657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,2560,0.05856621927685208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,3584,0.01612622208065457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,8192,0.011725333001878528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,10240,0.07967288626564874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,10240,0.01498311095767551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,5120,0.05087733268737793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,128,0.008115555677149031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,3584,0.009880888793203566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,1024,0.0042444442709287005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,3584,0.007996444072988298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,1536,0.010608889162540436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,128,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,7168,0.028235554695129395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,1024,0.004044444610675176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,2048,0.011681777735551199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,128,0.0029288888391521242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,6144,0.03672000103526645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,3072,0.0069271110826068455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,512,0.004541333350870344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,3072,0.014608888162506951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,16384,0.014102222190962898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,1536,0.016512889001104567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,6144,0.010509333676762052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,10240,0.04188088907135857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,7168,0.12237332926856147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,16384,0.006523555351628198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,768,0.006218666831652324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,2048,0.03933688998222351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,4096,0.02647466626432207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,1536,0.013173333472675748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,10240,0.012651555240154266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,10240,0.016235555211702984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,32,0.0031902222997612427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,1024,0.005670222143332164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,16384,0.03918222255176968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,65536,0.09050133493211533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,3072,0.007496889266702864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,2560,0.024561777710914612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,128,0.004086222085687849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,128,0.005396444350481033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,32,0.0034622223012977173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,8192,0.015479111009173922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,8192,0.02968177861637539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,5120,0.008544888761308458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,256,0.006496888895829518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,16384,0.03894044293297662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,6144,0.10383288727866279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,10240,0.09595733218722874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,64,0.0054560001525614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,2560,0.005912000106440649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,5120,0.16200355688730875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,5120,0.007186666958861881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,1024,0.023969777756267126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,3584,0.018048889107174344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,3584,0.007516444557242923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,7168,0.01977777812216017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,6144,0.03780444463094076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,768,0.005436444448100195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,64,0.0052897776994440295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,4096,0.015191111299726697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,2048,0.03770666652255588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,1536,0.005490666462315454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,5120,2.1785608927408853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,512,0.006280888699822956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,128,0.003536000019974179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,2560,0.006633777585294511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,768,0.012518222133318583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,4096,0.11220267083909775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,65536,0.19122221734788683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,768,0.03255555695957608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,4096,0.016372443901167977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,5120,0.022152889106008742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,128,0.003170666595300039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,3584,0.017647999856207106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,512,0.007309333317809635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,7168,0.021207110749350652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,128,0.005704000178310607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,2560,0.006032889087994893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,512,0.005623999983072281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,2560,0.008614222208658854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,2560,0.01110222190618515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,32,0.0024071110205517872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,128,0.05025688807169596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,16384,0.02036977807680766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,1536,0.006511999915043513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,7168,0.25712890095180935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,6144,0.29511910014682347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,6144,0.026848889059490625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,8192,0.08277422189712524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,16384,0.12420977486504449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,4096,0.018749333090252347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,128,0.008564444051848518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,512,0.0967831081814236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,8192,0.022620444496472675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,4096,0.011333333121405708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,3072,0.007648888561460707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,65536,0.10339555475446914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,7168,0.013754666679435305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,512,0.005476444545719359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,6144,0.012076444096035428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,4096,0.010285333626800114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,6144,0.010023111270533668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,5120,0.023898666103680927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,2048,0.16168622175852457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,2560,0.007623111208279927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,5120,0.014377777775128683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,1024,0.012661333713266583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,1536,0.09035911162694295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,1536,0.015470221638679504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,256,0.00463733325401942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,5120,0.909971555074056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,128,0.011752000285519494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,2048,0.015557333827018738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,2048,0.008163555628723567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,768,0.007195555501514011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,16384,0.2210728857252333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,2048,0.01757866640885671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,512,0.004577777865860197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,16384,0.06632088952594332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,6144,0.10859821902381049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,7168,0.07004000080956353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,6144,0.026473777161704168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,2560,0.007391110890441471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,5120,0.09715466366873847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,512,0.006875555548402999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,1536,0.01181777815024058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,7168,0.0779013302591112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,128,0.0032257777121331957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,256,0.004403555558787452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,128,0.006791111081838608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,7168,0.1338355541229248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,8192,0.024767110745112102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,2048,0.016009777784347534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,8192,0.02209866709179348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,12288,0.07968533039093018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,2048,0.006739555547634761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,3072,0.024333332975705464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,1024,0.10031288862228394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,768,0.00370755543311437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,256,0.005703111075692707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,1024,0.014746665954589844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,64,0.005111999809741974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,7168,0.014459555347760519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,6144,0.009511111511124505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,3584,0.0177511109246148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,3072,0.01000711073478063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,7168,0.12133955955505371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,2560,0.01815200017558204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,2048,0.018389332625601027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,6144,0.018458666072951425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,3584,0.007631111476156447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,2048,0.020270221763186984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,12288,0.03924533393647935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,128,0.010394666757848527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,256,0.004570666700601578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,7168,0.06663288672765096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,2048,0.007685333490371704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,10240,0.01642666591538323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,512,0.006751110984219446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,2048,0.00758133331934611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,256,0.0036231109665499795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,12288,0.01939644416173299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,5120,0.014152000347773233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,2560,0.028745777077145044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,32,0.0053288886944452924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,512,0.005133333305517833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,12288,0.9960346221923828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,2048,0.030573334958818223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,64,0.005067555440796746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,3072,0.01792711185084449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,128,0.005131555514203178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,32,0.08575822247399224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,65536,0.14336000548468694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,8192,0.02390488816632165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,128,0.005320888840489917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,7168,0.00961688905954361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,3072,0.036906666225857206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,512,0.006709333509206772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,10240,0.013542222479979197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,3072,0.0063822223908371395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,12288,0.1606986655129327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,12288,0.010868444210953183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,6144,0.0058453331391016645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,64,0.004570666700601578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,256,0.003168888803985384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,2560,0.018904889623324077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,2048,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,512,0.005070222334729301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,2560,0.016905777984195285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,10240,0.025732444392310247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,1024,0.010151111417346531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,128,0.0029360000044107437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,512,0.005819555371999741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,4096,0.01646222174167633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,1024,0.010253333383136326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,1536,0.0046942221621672315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,3072,0.009497777455382878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,5120,0.018917333748605516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,128,0.006244444598754247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,16384,0.5421848826938206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,7168,0.041201776928371854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,128,0.0028373334142896864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,5120,0.17615644137064615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,10240,0.21056089136335585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,8192,0.011681777735551199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,32,0.002692444456948174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,5120,0.015114666687117683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,16384,0.3154711193508572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,12288,0.01128088848458396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,3584,0.7101928922865125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,3584,0.016523554921150208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,32,0.02641422218746609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,3584,0.020671111014154222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,2560,0.20192178090413412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,65536,0.02829066581196255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,1024,0.006440889090299606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,2560,0.3491822348700629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,32,0.003428444473279847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,65536,0.6755581961737739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,3072,0.015781333049138386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,1536,0.007316444483068254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,128,0.03401688734690348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,64,0.003010666618744532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,4096,0.038783109850353666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,65536,0.2580080032348633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,6144,0.07304444577958849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,128,0.0032506665835777917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,12288,0.007140444384680853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,128,0.006333333336644703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,32,0.006755555669466655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,12288,0.06627466943528917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,6144,0.012780444489585029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,1536,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,8192,0.018072888255119324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,65536,0.1367502212524414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,768,0.005061333378156026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,1024,0.005943111247486538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,1024,0.008622222476535374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,8192,0.0084915558497111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,768,0.0069511110583941145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,256,0.0032479998966058097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,2048,0.008291555775536431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,65536,0.09025955862469143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,4096,0.09232978026072185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,6144,0.026390223039521113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,2048,0.005501333210203383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,3072,0.008197333249780867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,64,0.003228444399105178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,1536,0.007726222276687622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,5120,0.01722400055991279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,256,0.0031439999325407874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,6144,0.060843553808000356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,768,0.013917333549923368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,16384,0.02251111136542426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,6144,0.08525066905551487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,8192,0.030171556605233088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,12288,0.019685332973798115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,2560,0.005977777971161737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,32,0.010664888554149203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,65536,1.9657654232449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,5120,0.013213333156373767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,1536,0.007058666812049017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,7168,0.03930222325854831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,1536,0.02548444436656104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,7168,0.18886044290330675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,3072,0.02273155583275689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,1024,0.006761777732107375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,128,0.006716444260544247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,2048,0.01331555595000585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,65536,0.36763644218444824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,768,0.003584889074166616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,12288,0.43449065420362687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,6144,0.009902222288979424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,2560,0.007393777370452881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,2048,0.008685333033402761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,16384,0.04898044466972351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,512,0.004942222187916438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,7168,0.04965066578653124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,128,0.0032631111227803757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,10240,0.007107555866241455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,12288,0.04687111245261299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,10240,0.1837599939770169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,12288,0.0536426670021481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,12288,0.007920888562997183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,1536,0.007198221981525421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,3072,0.016753777861595154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,6144,0.00685066646999783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,768,0.00452977791428566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,3072,0.01036800030204985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,512,0.004056888735956616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,5120,0.0094373333785269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,65536,0.012650666965378655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,4096,0.009523555636405945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,1024,0.005390222287840313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,6144,0.016174222032229107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,1536,0.005758222192525864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,2048,0.009533333281675974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,768,0.005222222043408289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,128,0.005021333280536863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,5120,0.16961422231462267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,8192,0.09277244408925374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,16384,0.030731555488374498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,768,0.00360622215602133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,6144,0.051628443929884166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,10240,0.03495733274353875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,12288,0.8715368906656901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,256,0.006776000062624614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,8192,0.09583999713261922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,7168,0.021320889393488567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,6144,0.009701333112186855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,8192,0.031535999642478094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,1536,0.07849244276682536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,2560,0.005680888891220093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,7168,0.015992888145976596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,7168,0.14096355438232422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,128,0.0058453331391016645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,1024,0.004262222184075249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,256,0.006760889043410619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,7168,0.006051555689838197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,512,0.0058613332609335584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,16384,3.45489501953125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,8192,0.03628266519970364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,5120,0.0101742222905159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,8192,0.015090665883488126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,128,0.02847111225128174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,5120,0.01651822196112739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,12288,0.013092444174819522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,4096,0.02070488863521152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,3072,0.011016888750924004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,768,0.012655999925401477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,10240,0.09692266914579604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,2048,0.01089599976936976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,12288,0.09438133239746094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,32,0.005955555372767978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,128,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,64,0.0028248888750871024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,64,0.0052435555391841466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,4096,0.01146933353609509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,1024,0.014269333746698169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,2048,0.0052791109515561005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,1536,0.0069155556460221606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,1024,0.009383111364311641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,65536,0.13253422578175864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,8192,0.04952355557017856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,65536,0.15153333875868055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,512,0.010406222608354356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,3072,0.012705778082211813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,64,0.004703999807437261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,6144,0.013890667094124688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,3072,0.012268444730175866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,2048,0.006542222367392645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,512,0.005767111149099138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,768,0.008943999807039896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,3584,0.0240666667620341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,65536,0.271708435482449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,3584,0.03587111168437534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,16384,0.05405866768625048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,65536,0.5909057723151313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,1536,0.006108444597986009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,3584,0.017097777790493436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,768,0.009332444104883406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,8192,0.02156266735659705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,128,0.004551110996140374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,6144,0.018187556001875136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,64,0.006824888702895906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,2048,0.012177777787049612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,32,0.00272088891102208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,3072,0.12095377180311416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,6144,0.0158897770775689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,10240,0.026086222794320848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,512,0.005893333504597346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,3584,0.010696888797812991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,4096,0.016773333152135212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,1024,0.03374222252104018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,8192,0.40773865911695695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,6144,0.018699554933442008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,1536,0.005763555566469829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,12288,0.05423733260896471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,256,0.059002664354112416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,7168,0.01368533323208491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,3584,0.007877333296669854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,5120,0.01641155613793267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,512,0.016470222009552848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,65536,1.5109679963853624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,10240,0.05088622371355692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,12288,0.08629333310657078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,3584,0.020087111327383254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,64,0.0035866668654812705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,32,0.005903999838564131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,12288,0.027653333213594224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,4096,0.07170222202936809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,64,0.0033608888172441055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,16384,0.041407111618253924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,64,0.0075706665714581805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,8192,0.022279111875428095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,4096,0.06557777855131361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,32,0.013831111292044321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,256,0.02753511071205139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,2048,0.00565422202150027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,2560,0.009509333305888707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,1536,0.20876888434092203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,6144,0.04832355512513054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,128,0.004302222281694412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,7168,0.041688889265060425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,6144,0.01391377713945177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,10240,0.055664002895355225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,65536,0.16304977734883627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,8192,0.13089866108364528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,5120,0.006611555400821898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,5120,0.009700444837411245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,5120,0.013218666944238873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,8192,0.014020444618331062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,5120,0.030534221066368952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,256,0.004938666605287128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,12288,0.04833955566088358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,64,0.003592888928121991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,6144,0.02441155579355028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,65536,0.0735644433233473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,5120,0.00870311094654931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,12288,0.021078222327762183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,10240,0.04788177874353197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,7168,0.02940888868437873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,65536,0.5150515768263075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,8192,0.025476444098684523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,4096,0.020602666669421725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,3584,0.017045333981513977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,256,0.004202666795916027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,12288,0.04163644380039639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,16384,0.06196710798475477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,2048,0.0053280000057485365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,1024,0.00700711127784517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,2560,0.091030220190684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,3584,0.022055110997623865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,1024,0.005424889011515512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,3072,0.019660444723235238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,1024,0.0037502224246660867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,64,0.005180444568395615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,7168,0.026148445076412622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,2560,0.006024000131421619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,7168,0.15644799338446722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,768,0.006095110956165526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,2560,0.01203555530971951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,8192,0.05662044551637438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,1024,0.012159999873903064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,2048,0.005931555396980709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,2560,0.014372444815105863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,2048,0.015657777587572735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,3584,0.01016977760526869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,6144,0.008360889222886827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,5120,0.010667555862002902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,2560,0.011834666960769229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,3584,0.1600737836625841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,32,0.003050666716363695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,3072,0.0064382221963670515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,5120,0.02149066660139296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,3584,0.02032177812523312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,32,0.0052684446175893145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,64,0.0026782221264309357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,8192,0.00928177767329746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,768,0.003887999802827835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,5120,0.048215111096700035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,1536,0.3258204460144043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,3072,0.008519111408127679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,5120,0.012282667060693106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,2560,0.04320799973275927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,256,0.011567110816637674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,7168,0.01423555612564087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,4096,0.0932391087214152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,256,0.0033582221302721235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,512,0.0036657779581016968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,4096,0.00853777759604984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,10240,0.04068888889418708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,128,0.007151111132568783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,1536,0.005532444351249271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,65536,0.2176133394241333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,2048,0.007648888561460707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,8192,0.10339999861187404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,128,0.0027662221756246355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,7168,0.02750844425625271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,3584,0.028749333487616643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,2048,0.007035555938879649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,3584,0.41297954983181423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,3584,0.01793866687350803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,32,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,8192,0.03834133346875509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,512,0.005588444570700328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,12288,0.9105929268731011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,3584,0.23758133252461752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,10240,0.014532445205582513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,3072,0.010832000109884473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,5120,0.03544888893763224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,7168,0.021735111872355144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,256,0.007142222589916653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,10240,0.021521778570281133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,768,0.008384000096056197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,32,0.008573333422342936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,768,0.005168888717889786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,3072,0.009561777942710454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,256,0.015085332923465304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,3072,0.026555554734336004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,5120,0.045518222782346934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,3072,0.00941422250535753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,3072,0.008777778181764815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,3584,0.021507556239763897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,2048,0.006130666782458623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,65536,0.21293511655595568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,1536,0.028487109475665625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,12288,0.030887110365761652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,64,0.006170666466156642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,10240,0.038719111018710665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,1536,0.006519111080302133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,4096,0.008176888856622908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,6144,0.00629688882165485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,1536,0.21697955661349824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,5120,0.009474666582213508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,2560,0.013277333643701343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,16384,0.02520622147454156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,8192,0.03695733348528544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,10240,0.06056444512473213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,8192,0.13961866166856554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,1536,0.005797333187527127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,2560,0.011139555937714048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,2048,0.09648178021113078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,16384,14.165636698404947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,4096,0.06523378027809991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,65536,0.10619199938244289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,16384,0.025331555141343012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,12288,0.020961778031455144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,6144,0.01427288850148519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,8192,0.03186666634347703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,2560,0.01146666705608368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,64,0.005877333382765452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,16384,0.022464889619085524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,7168,0.018072888255119324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,8192,0.05625066492292616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,1536,0.005345777918895085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,16384,0.017831999394628737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,32,0.003156444471743372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,6144,0.006346666564544042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,3072,0.007484444313579136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,32,0.0026088888860411118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,256,0.012735111018021902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,768,0.00599644457300504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,10240,0.07983199755350749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,2048,0.010822222464614444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,10240,0.01808088852299584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,10240,0.015254222684436373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,4096,0.40383466084798175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,64,0.0031102223114834893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,8192,0.06278310881720649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,12288,0.017312889297803242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,65536,0.06928266419304742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,768,0.009725333087974125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,256,0.0028657778683635923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,512,0.06315910816192627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,2560,0.006401777681377199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,128,0.0028764444092909494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,7168,0.03126488791571723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,512,0.006062222023804982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,128,0.0033288889875014624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,8192,0.014179555906189812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,2048,0.012327999704413943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,128,0.002597333242495855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,12288,0.023971555961502924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,768,0.004823110997676849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,5120,0.022062222162882488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,1536,0.007908444437715743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,512,0.003236444460021125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,32,0.006523555351628198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,1024,0.008639111287064022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,256,0.05751644240485298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,1536,0.00519466648499171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,6144,0.13277333312564427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,7168,0.030679110023710463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,65536,0.07965066697862414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,2560,0.0064284445510970215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,768,0.004281777888536453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,16384,0.040799111127853394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,64,0.0026168889469570587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,4096,1.657499525282118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,8192,0.9252711402045356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,128,0.005370666583379109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,3584,0.024988444315062627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,3072,0.012498666842778524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,65536,2.673110326131185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,2560,0.009628444082207149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,256,0.003960888832807541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,1536,0.010234666367371878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,512,0.02737155556678772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,7168,0.5008666780259874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,8192,13.700964185926649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,3584,0.012327111429638334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,64,0.002512000087234709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,6144,0.04131733377774557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,5120,0.013929777675204806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,8192,0.07288444704479642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,3072,0.028065777487225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,4096,0.010022222167915767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,6144,0.018823110395007663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,3072,0.016928889685206942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,128,0.005114666703674528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,64,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,5120,0.012758221891191272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,12288,0.024061333802011278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,7168,1.1785635418362088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,768,0.011115555961926779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,64,0.0053395554423332214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,768,0.006992888947327931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,32,0.005524444497293896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,128,0.004558222161398994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,128,0.0047475554876857335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,256,0.007149332927332983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,512,0.025099555651346844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,7168,0.032368001010682844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,3584,0.017676444517241586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,16384,0.6269511116875542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,4096,0.04038844505945841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,4096,0.012900444368521372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,4096,0.02308977809217241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,1024,0.004594666676388847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,6144,0.056567112604777016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,5120,0.1013608906004164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,128,0.00286311118139161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,12288,0.009642666412724389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,2560,0.006567111031876669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,256,0.005656888915432825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,16384,0.02292088833120134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,1536,0.009130666653315226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,128,0.005192888693677055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,1024,0.05228622092141045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,32,0.006664000037643645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,768,0.008041777544551426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,65536,0.07299377520879109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,3072,0.011114666859308878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,65536,0.024696888195143804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,512,0.006144888699054718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,7168,0.01715555621518029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,16384,0.05169422096676297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,64,0.00507733349998792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,12288,0.011759999725553723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,8192,0.034607112407684326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,6144,0.019112888309690688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,2560,0.007439110842016008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,1536,0.010487111078368293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,16384,0.006140444427728653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,4096,0.011384000380833944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,128,0.005158222383923001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,768,0.007501333124107785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,12288,0.02539199921819899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,7168,0.020989333589871723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,512,0.005756444401211209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,32,0.0031235555393828284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,12288,0.014921777778201632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,65536,1.3761680391099718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,10240,0.05035555693838331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,768,0.004171555654870139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,768,0.005967111223273807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,512,0.006893333461549547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,10240,0.01127022256453832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,1536,0.008085333638721043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,2048,0.013088888592190213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,8192,0.008692444198661381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,1024,0.006491555521885554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,3584,0.0074480002125104266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,8192,0.006786666810512543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,2560,0.01462222304609087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,2048,0.005930666708283954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,65536,0.06557244724697538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,2048,0.02698488864633772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,7168,0.016564443707466125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,5120,0.007858666280905405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,512,0.005501333210203383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,3584,0.008768888811270395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,256,0.006918222126033571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,16384,0.04707644383112589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,4096,0.010959111154079437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,512,0.00592533333433999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,2560,0.19107466273837617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,2560,0.10038577848010594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,6144,0.01405422223938836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,32,0.006960888703664144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,16384,0.054979556136661105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,128,0.005117333183685939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,6144,0.14593777391645643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,10240,0.01648533344268799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,7168,0.007282666862010956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,8192,0.01512888901763492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,3072,0.06057777669694689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,2560,0.00775999989774492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,12288,0.016698666744761996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,1024,0.1961359977722168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,5120,0.02646044393380483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,12288,0.045422222879197865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,65536,0.08964711427688599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,512,0.00775999989774492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,8192,0.029321776496039495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,512,0.004729777574539185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,4096,0.006811555474996567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,3584,0.014667555689811707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,5120,0.0064284445510970215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,1536,0.007420444654093848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,1536,0.415483554204305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,65536,0.08382488621605767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,8192,0.02479733361138238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,256,0.005762666463851929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,8192,0.16025066375732422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,12288,0.3746151129404704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,16384,0.4402471118503147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,1536,0.15170489417182073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,1024,0.00545866663257281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,64,0.0029599999801980127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,6144,9.911989000108507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,6144,0.01035288886891471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,256,0.0035439998739295532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,12288,0.007557333343558841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,128,0.007912000020345053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,5120,0.00924800005224016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,1024,0.025582222474945918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,10240,0.26044265429178876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,128,0.004390222330888112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,5120,0.2778106795416938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,5120,0.11141333315107559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,1536,0.006436444405052397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,3072,0.014183999763594734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,32,0.004937777916590373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,768,0.07257422473695543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,2048,0.005626666463083691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,12288,0.07230933507283528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,2560,0.01161688897344801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,512,0.0055440002017551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,3072,0.01648888819747501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,256,0.005296000176005893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,4096,0.009319999979601966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,10240,0.023123555713229712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,65536,3.240943060980903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,32,0.0041128889554076726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,128,0.0033608888172441055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,12288,0.02239022155602773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,6144,0.009653333160612319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,3584,0.014389332797792224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,256,0.003649777836269803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,16384,0.0916266640027364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,1536,0.007259555160999298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,12288,0.32558488845825195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,16384,0.052889777554406055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,12288,0.08882044421301948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,256,0.00926577796538671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,128,0.004115555435419083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,512,0.005146666533417172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,12288,0.014110222458839417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,3072,0.0074942219588491656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,1536,0.01978933314482371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,64,0.002573333266708586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,6144,0.036644445525275335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,3072,0.006509333435032103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,4096,0.008280889027648503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,512,0.003535110917356279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,2560,0.021856889128684998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,4096,0.006480888773997624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,1536,0.06564444303512573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,256,0.004576000074545543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,256,0.005370666583379109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,2560,0.023924445112546284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,10240,0.183695117632548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,2048,0.1020817756652832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,1536,0.056270221869150795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,5120,0.012078222301271228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,64,0.002471111093958219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,16384,0.05182222194141812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,64,0.005126222140259213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,8192,0.026320889592170715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,5120,0.017589333984586928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,12288,0.012120000190205045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,256,0.0034373334298531213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,7168,0.010178666975763109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,1536,0.009226666556464301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,64,0.005034666508436203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,64,0.002607999990383784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,4096,0.014346667461925082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,10240,0.04461244410938687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,1536,0.013946667313575745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,8192,0.006513777706358168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,2048,0.005517333332035277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,65536,0.04278488953908285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,3584,0.03973244296179877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,5120,0.19235200352138945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,128,0.0036862223512596558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,1024,0.004371555729044808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,512,0.0058311112225055695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,10240,0.1269635491900974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,1024,0.004580444345871608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,2048,0.007384888827800751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,65536,0.10893244213528103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,2560,0.006299555715587404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,10240,0.036350223753187395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,1024,0.006118222243256039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,6144,0.13459732797410753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,128,0.007636444436179266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,32,0.002760888801680671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,12288,0.3910097810957167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,512,0.003351111171974076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,7168,0.01607911123169793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,2560,0.005955555372767978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,7168,0.017551110850440133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,6144,0.4506524403889974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,128,0.004129777765936322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,256,0.0052879999081293745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,16384,0.01945066617594825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,64,0.0061048890153567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,3584,0.007048889166778988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,6144,0.012860444684823355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,8192,0.02259911100069682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,64,0.0027884443600972495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,1024,0.08322844240400526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,768,0.0037333332002162933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,2048,0.032708442873424955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,256,0.01012711144155926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,3072,0.3176328870985243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,12288,0.01980799933274587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,6144,0.2507750988006592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,6144,0.010870222416188983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,16384,0.00924711094962226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,256,0.015232000086042615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,4096,0.032462222708596125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,12288,0.4024799929724799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,256,0.003597333199448056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,2048,0.007804444266690149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,16384,0.025456888808144465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,4096,0.04025777843263414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,3072,0.10002488560146755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,6144,0.010042666561073726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,3072,0.009233777721722921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,3584,0.053520000643200345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,1024,0.003927111211750242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,12288,0.023992889457278784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,16384,0.06720177994834052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,64,0.002704888789190186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,128,0.005423999908897612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,10240,0.022479999396536086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,5120,0.007318221860461765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,512,0.005867555737495422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,5120,0.026202667090627883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,3584,0.011982222398122152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,2048,0.006005333529578314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,6144,0.005865777946180767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,2560,0.013732444081041547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,768,0.3059635427263048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,4096,0.011061333119869232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,32,0.0025964443468385274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,4096,0.019725332657496136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,3584,0.05658044417699178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,1536,0.04296888907750448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,8192,0.023914666639433965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,128,0.004959110998445087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,512,0.006088888893524806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,6144,0.03513155712021722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,5120,0.0090515555606948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,4096,0.023394667439990576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,10240,0.06636799706353082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,512,0.028419554233551025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,7168,0.1853546698888143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,2560,0.05432088838683235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,1024,0.008995555341243744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,512,0.010453333457310995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,10240,0.014344000154071383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,3072,0.006724444528420766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,65536,0.06854222218195598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,5120,0.011126221881972419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,8192,0.13301777839660645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,128,0.0063946665161185795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,512,0.011727111207114326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,1536,0.5793369081285265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,65536,0.11712177594502766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,256,0.006755555669466655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,2560,0.00848355558183458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,8192,0.017130666308932833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,8192,0.012344000240166983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,10240,0.0362702210744222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,768,0.006784889019197888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,128,0.00996088898844189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,64,0.06301244762208727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,5120,0.02375288969940609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,2048,0.01236177815331353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,2560,0.019917334119478863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,3584,0.023039110832744177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,2048,0.006660444455014334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,2048,0.01685422162214915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,512,0.0034959999223550162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,5120,0.015691555208630033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,7168,0.023904000719388325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,2560,0.006497777584526274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,3584,0.02472977836926778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,1024,0.007988444632954067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,65536,0.07663377788331774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,3584,0.009321778184837764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,32,0.00309244439833694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,1536,0.04865511258443197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,10240,0.008225777910815345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,2560,0.010753778119881948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,2560,0.04428799947102865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,128,0.003006222140457895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,4096,0.00703022215101454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,3072,0.012624888784355588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,4096,0.023039110832744177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,4096,0.014398222168286642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,3072,0.008117333054542542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,512,0.003458666718668408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,1024,0.007687110867765214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,5120,0.013957333233621387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,128,0.004335999902751711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,8192,0.06627644432915582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,3584,0.00906399968597624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,65536,0.1047475536664327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,2048,0.008016000191370646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,2048,0.0057946667075157166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,4096,0.007912000020345053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,12288,0.01513777838812934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,3584,0.0107342220014996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,2560,0.05781866444481743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,512,0.003022222262289789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,10240,0.07909688684675428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,128,0.009010666774378883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,65536,0.31529333856370717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,768,0.019107555349667866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,16384,0.025791999366548326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,64,0.014682667122946845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,7168,0.0337759984864129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,2048,0.11801066663530137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,2048,0.009849777652157677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,512,0.010034666293197209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,3072,0.011776888536082374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,12288,0.009869332942697737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,16384,0.023733332753181458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,12288,0.08569333288404678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,512,0.003539555602603488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,3584,0.008283555507659912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,12288,0.016855110724767048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,128,0.0025804444319672054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,12288,0.2987146642473009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,256,0.00378222225440873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,4096,0.020052444603708055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,6144,0.01787911189926995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,64,0.0027751111321979095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,64,0.024494222468800012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,65536,0.21607289049360487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,32,0.007500444849332173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,1024,0.004406222038798862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,64,0.002663111107216941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,128,0.033504890071021184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,64,0.004442666553788715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,12288,0.05119555526309543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,4096,0.00777422222826216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,32,0.002667555585503578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,2048,0.00832177781396442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,3072,0.008488000267081791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,32,0.005130666825506422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,7168,0.011590222517649332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,7168,0.013030221892727746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,1536,0.011247999966144562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,1024,0.0418471097946167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,128,0.00847999999920527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,1024,0.007556444240940942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,2560,0.009203555683294931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,512,0.006871999965773688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,1024,0.006063111126422882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,5120,0.009509333305888707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,4096,0.02800977892345852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,128,0.005151999907361137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,512,0.003335111050142182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,128,0.011344888971911537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,3584,0.018026666508780587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,8192,0.014793778459231058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,2560,0.015094222293959724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,5120,0.043770667579438954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,7168,0.5467564264933268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,7168,0.04468355576197306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,32,0.0026577777332729767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,4096,0.010509333676762052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,2048,0.008753777378135258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,1536,0.006051555689838197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,5120,0.011018666956159802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,7168,0.019271110494931538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,3584,0.021168889270888433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,32,0.03162844313515557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,2048,0.0162808895111084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,65536,0.9110000398423937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,128,0.008563555777072906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,3584,0.008768888811270395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,256,0.002834666727317704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,8192,0.011368888947698804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,10240,0.0226248883538776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,4096,0.013210666676362356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,16384,0.32227200931972927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,2048,0.01387466655837165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,768,0.13457332717047796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,2560,0.00815466625822915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,65536,0.020808888806237113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,8192,0.09981866677602132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,3072,0.008535111116038429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,3072,0.03233244352870517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,4096,0.009069333473841349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,6144,0.014642667439248828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,3072,0.006572444405820634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,768,0.009167999857001835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,5120,0.012476444244384766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,2048,0.008344888687133789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,6144,0.026615111364258662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,2048,0.09436533186170791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,256,0.0053386667536364655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,16384,0.03717866539955139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,5120,0.13818221622043186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,6144,0.04987200101216634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,10240,0.025450666745503742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,1024,0.021032000581423443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,16384,0.15028356181250677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,5120,0.021002666817771062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,768,0.005950222412745158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,6144,0.20796710915035674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,12288,0.10300622383753459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,768,0.005119110975000593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,8192,0.02838488750987583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,10240,0.01940711173746321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,4096,0.011379555695586734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,512,0.004574222283230888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,3584,0.015598222613334656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,3584,0.01662133302953508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,2048,0.009166222479608323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,5120,0.2925848960876465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,4096,0.00870488915178511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,65536,0.02221155497762892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,5120,0.031410667631361223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,512,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,2560,0.006513777706358168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,32,0.007806222471925948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,768,0.007286222444640265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,2048,0.006578666468461354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,768,0.0269768883784612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,10240,0.016262221667501662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,3072,0.008457777400811514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,2048,0.008486222061845992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,2560,0.007153777612580194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,4096,0.006372444331645966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,512,0.009640889035330879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,5120,0.008644444247086843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,3072,0.012828444441159567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,5120,0.009342222577995723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,512,0.0047146665553251905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,2048,0.30035556687249076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,2560,0.058655997117360435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,256,0.0047351109484831495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,2560,0.04862844281726413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,12288,0.020399111840460036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,5120,0.030288888348473444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,5120,0.021551999780866835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,64,0.002687999978661537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,10240,0.04916977882385254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,1024,0.004769777672158347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,64,0.012061333490742577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,65536,0.1609333356221517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,6144,0.08078044652938843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,4096,0.018615111708641052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,64,0.0025333333760499954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,2048,0.018349332941903006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,1024,0.032771554258134626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,12288,0.21545778380499947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,16384,0.9455849329630533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,512,0.005532444351249271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,512,0.006384000182151794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,2560,0.01125955581665039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,6144,0.017624000708262127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,768,0.005670222143332164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,768,0.0047626665068997275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,3584,0.017983999517228868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,4096,0.017416889468828838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,65536,0.5956995752122667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,4096,0.019706666469573975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,4096,0.00817599975400501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,6144,0.26796089278327095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,10240,0.03842844565709432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,3584,0.01037688884470198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,32,0.0028417776856157514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,4096,0.009122666385438707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,3072,0.01809244520134396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,256,0.0029528888149393927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,8192,0.022336888644430373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,1024,0.0052355556852287715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,16384,0.03141955534617106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,3072,0.09270399808883667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,12288,0.019692444139056735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,1536,0.006017777654859755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,1024,0.010670222342014313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,512,0.0032906666811969546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,6144,0.01128888875246048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,128,0.005416000054942236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,16384,0.021401777863502502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,8192,0.029338667790095013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,128,0.0027475555737813315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,65536,0.2532142268286811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,2560,0.013581333061059317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,3584,0.007316444483068254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,10240,0.032617777585983276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,128,0.018449778358141582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,4096,0.06349066893259685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,2048,0.007635555333561367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,2048,0.014136888914638095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,512,0.00408266650305854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,512,0.005610666755172942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,5120,0.020289777053727042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,3584,0.008108444511890411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,768,0.006210666563775804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,65536,0.0664586689737108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,2048,0.02659999993112352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,2560,0.026867555247412786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,128,0.0027626665929953256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,16384,0.02130311065249973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,8192,0.038077334562937416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,2560,0.007182222273614671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,768,0.008032889001899296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,3072,0.015407999356587728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,2048,0.004944889081848992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,16384,0.04873955580923292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,3584,0.008817777865462834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,12288,0.012821333275900947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,768,0.0036115555299652945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,5120,0.023154666026433308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,7168,0.029618667231665716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,1024,0.006974221931563483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,256,0.009750222166379293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,128,0.00572622236278322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,2560,0.007267555428875818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,7168,0.13141866525014242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,3584,0.007683555285135905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,8192,0.2643502288394504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,256,0.0054551110499435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,64,0.002601777720782492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,16384,0.03976088762283325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,4096,0.011113777756690979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,64,0.003990222182538775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,65536,0.14044088787502712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,16384,0.020816889074113634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,1536,0.006503110958470239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,1536,0.0069368887278768755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,7168,0.013316444224781461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,3584,0.021350221501456365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,768,0.010209778116808997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,10240,0.04885333445337084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,3072,0.0144177774588267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,512,0.020474667350451153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,2560,0.01723377737734053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,1024,0.015523556205961438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,32,0.04202933443917168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,128,0.0047484445903036325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,10240,0.10281155506769817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,5120,0.08157422145207723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,4096,0.05857066975699531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,128,0.004121777911980947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,8192,0.00869777798652649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,1024,0.003989333493842019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,4096,0.8201413154602051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,3072,0.024122666981485155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,768,0.006322666588756774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,6144,0.04137333234151205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,256,0.004363555461168289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,2560,0.026241777671708003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,512,0.006867555694447623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,3072,0.014426666829321118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,65536,0.0800959997706943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,8192,0.025127111209763422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,5120,0.20796710915035674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,12288,0.03700177868207296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,2560,0.08379555410808986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,32,0.004964444372389051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,1024,0.004913777940803104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,3584,0.008201777935028076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,1536,0.03147644466824002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,1536,0.1112711164686415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,4096,0.00814933329820633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,6144,0.06245599852667915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,768,0.00611200018061532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,65536,0.018343110879262287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,512,0.0033973333322339584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,10240,0.035974221097098455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,3584,0.011919111013412476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,256,0.008607111043400234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,2048,0.008799999952316284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,3584,0.018166222506099276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,2048,0.010221333139472539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,64,0.0053280000057485365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,7168,0.006135111053784688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,64,0.04063466522428725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,1536,0.006110222389300664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,32,0.0036346668170558084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,65536,0.06217600239647759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,5120,0.0732497771581014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,2560,0.062033779091305204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,6144,0.013428444663683573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,3584,0.012073777616024017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,8192,6.430048200819228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,1536,0.1093484428193834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,3072,0.008962666822804345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,4096,0.034713778230879046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,128,0.007150222029950883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,3584,0.007538666327794393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,5120,0.012355555262830524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,1536,0.02887466549873352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,2560,0.00720177756415473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,3072,0.006762666834725275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,3072,0.13282310962677002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,2560,0.009226666556464301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,5120,0.011876444849703046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,1536,0.014204444156752693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,7168,1.0576480229695637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,768,0.007269333634111617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,6144,0.05886933538648817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,7168,0.06995999813079834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,6144,0.013234666652149625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,256,0.004777777940034866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,1536,0.011847111086050669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,65536,0.6666480170355903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,12288,0.022543110781245764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,128,0.002986666642957263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,128,0.004960000101062986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,2560,0.007057777709431118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,65536,5.270409054226346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,128,0.0037342223028341928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,5120,0.018532445033391316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,2560,0.037167999479505755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,10240,0.02184711065557268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,256,0.0034968890249729156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,16384,0.05621244510014852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,4096,0.007704888780911763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,512,0.00564444437623024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,16384,0.01166488892502255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,128,0.006527110934257507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,2048,0.011648889217111798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,65536,0.20039555761549208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,3072,1.2909155951605904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,256,0.008372444245550368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,768,0.019513777560657926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,5120,0.008778666456540426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,3584,0.00902133352226681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,65536,0.1274284389283922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,12288,0.2671804428100586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,16384,0.024580443898836773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,5120,0.03300889001952277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,4096,0.12822488943735758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,7168,0.04033599959479438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,12288,0.0461368891927931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,4096,0.024513777759340074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,5120,0.009443555441167619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,3072,0.3409244484371609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,8192,0.015138667490747241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,6144,0.026105778084860906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,3072,0.1084444456630283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,512,0.0033484444850020935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,32,0.005764444255166584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,5120,0.024138665861553613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,65536,0.3533182144165039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,2048,0.006270222365856171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,6144,0.05755644374423557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,16384,0.09027200274997288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,4096,0.006178666734033161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,12288,0.02132266594303979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,128,0.004042666819360522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,3584,0.023839111129442852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,12288,0.44058044751485187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,10240,0.01364177796575758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,4096,0.01631733278433482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,10240,0.03536444571283128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,1024,0.006543111056089401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,3072,0.009761778016885122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,4096,0.019578667150603402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,512,0.0068764446510208984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,512,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,65536,0.25046666463216144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,768,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,10240,0.013590222431553734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,1024,0.009679999616410997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,65536,0.03342399994532267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,10240,0.0964755548371209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,1536,0.015486222174432544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,16384,0.31222134166293675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,1536,0.006412444429265128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,7168,0.10314489073223537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,768,0.006219555520349079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,3072,0.011456888582971362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,1536,0.007253333098358578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,16384,0.020294222566816542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,65536,0.07847466733720568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,128,0.013821333646774292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,16384,0.10740888781017727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,256,0.004881777697139316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,7168,0.011010666688283285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,128,0.020390222469965618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,3584,0.01606399979856279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,65536,0.007430222299363878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,4096,0.04236355423927307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,128,0.0026595555245876312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,512,0.0032115555885765287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,3584,0.013182222015327878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,1024,0.006351999938488007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,1536,0.00647288892004225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,3584,0.01220266686545478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,10240,0.017115554875797696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,3584,0.00740533322095871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,4096,0.03321511215633816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,3584,0.0075084442893664045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,7168,0.01425866617096795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,32,0.005467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,65536,1.5177173614501953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,8192,1.6753004921807184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,512,0.060328887568579785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,32,0.003297777846455574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,1536,0.010793777803579966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,64,0.010505778094132742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,7168,0.008990222381220924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,1024,0.00434755575325754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,32,0.0038977778620190094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,128,0.009581333233250512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,256,0.005498666730191972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,6144,0.014108444253603617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,4096,0.018072000808186002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,12288,0.018798222144444782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,5120,0.013056889176368713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,128,0.004368888835112254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,256,0.0035822221802340615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,128,0.002806222273243798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,16384,3.0024435255262585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,1024,0.020968000094095867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,10240,0.014170666535695394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,3072,0.014645333091417948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,128,0.0045422220395671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,1024,0.009866666462686326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,6144,0.033222221665912204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,128,0.004719999929269155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,64,0.005723555468850666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,128,0.0038693332009845306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,3584,0.0392577780617608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,12288,0.16094311078389487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,128,0.010825778047243753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,3584,0.12128888236151801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,4096,0.01945333348380195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,128,0.004104888687531154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,32,0.0027422221998373666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,10240,0.3673582341935899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,4096,0.015923556354310777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,3072,0.007275555696752336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,5120,0.028804444604449805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,5120,0.021967111362351313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,12288,0.025216889050271776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,64,0.005130666825506422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,8192,0.013806222213639153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,10240,1.308068487379286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,128,0.004534222185611725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,4096,0.022262222237057153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,2560,0.012413333687517377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,10240,0.020798222886191476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,16384,0.02012355625629425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,768,0.006751110984219446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,65536,0.014945778581831189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,3072,0.0083155557513237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,128,0.004592888885074192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,2560,0.148181332482232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,768,0.020454222957293194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,64,0.004058666527271271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,3584,0.6538186603122288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,128,0.003855111284388436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,8192,0.012891555825869242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,2048,0.07519733243518405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,32,0.002749333365095986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,4096,0.014939554863505893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,5120,0.3344746695624457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,16384,0.057909336354997426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,1536,0.01331822243001726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,3072,0.00853066643079122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,768,0.007041778001520369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,1024,0.022088888618681166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,1536,0.006874666859706243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,12288,0.812968889872233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,1536,0.012667555775907306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,2048,0.008608889248636033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,65536,0.6058150927225748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,3584,0.017273777061038546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,3072,0.02308711078431871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,16384,0.8952488369411893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,3584,0.02604888876279195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,7168,0.04201777776082357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,4096,0.010122666756312052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,65536,0.07719110780292086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,1536,0.008776889079146916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,6144,0.1089608867963155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,2048,0.07234844234254625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,10240,0.3329226705763075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,64,0.0046977777447965415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,3584,0.008882666627566019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,4096,0.010272888673676385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,65536,0.08506755696402656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,2048,0.012910222013791403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,256,0.004287111262480418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,1024,0.017110221915774874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,256,0.005534222142563925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,2048,0.012998222476906247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,2560,0.19080266687605116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,1024,0.005459555735190709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,12288,0.0394755568769243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,4096,0.011044444309340583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,16384,0.09387289153205024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,768,0.0037724446091387006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,64,0.003176888864901331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,1536,0.004676444249020683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,1024,0.00887111077706019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,1024,0.005423999908897612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,16384,0.06266755527920194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,3584,0.009954666925801171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,2560,0.008631111019187504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,256,0.0031724443866146933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,1536,0.0070168889231152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,1024,0.005365333209435145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,12288,0.0663964417245653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,65536,0.15247110525767008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,8192,0.03222666515244378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,128,0.005158222383923001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,1024,0.005743999862008625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,128,0.0031146665828095544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,8192,0.01420177850458357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,12288,0.02126844392882453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,768,0.006628444625271692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,3072,0.009005332986513773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,7168,0.028928889168633357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,512,0.016452444924248587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,2048,0.008590222232871586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,7168,0.017120889491505094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,768,0.005566222386227713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,4096,0.010499555203649731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,5120,0.021914665897687275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,65536,0.20238844553629556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,512,0.00342755557762252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,5120,0.022094221578703985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,4096,0.02360711163944668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,2048,0.011095110740926532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,65536,0.12189422713385688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,16384,0.0210817770825492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,64,0.0026462222966882917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,1536,0.017673777209387887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,8192,0.01148799972401725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,4096,0.009023110899660323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,3072,0.009905777871608734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,8192,0.1778479946984185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,1024,0.020058666666348774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,128,0.0030657777355776895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,128,0.003128000017669466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,7168,0.12724088297949895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,3072,0.007173333730962541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,256,0.08703821897506714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,768,0.004535111288229625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,65536,0.17670399612850615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,65536,0.11343822214338516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,4096,0.0064533332155810455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,2048,0.005543111099137201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,6144,0.013266666895813413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,65536,0.14470222261216906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,65536,0.11938844786749946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,2048,0.0064186664919058485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,2560,0.008929777476522658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,4096,0.008900444540712569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,12288,0.02093688812520769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,512,0.0058204444746176405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,8192,0.019662222928471036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,2560,0.01835466590192583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,3584,0.006765333314736684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,5120,0.029771556456883747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,16384,0.04368355539109972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,512,0.006102222121424145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,1024,0.2907093365987142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,512,0.006156444549560547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,128,0.010278222461541494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,64,0.0033893332713180115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,8192,0.046478221813837685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,65536,20.251790364583332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,256,0.013993778162532382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,1024,0.010289777484205034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,8192,0.033450666401121355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,512,0.005897777775923411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,3072,0.009100444614887238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,3584,0.006230222268237009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,1536,0.005917333480384614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,4096,0.019122666782803006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,32,0.003573333223660787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,768,0.017927999297777813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,10240,0.013849777479966482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,10240,0.018752889500723947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,512,0.0035093331502543557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,64,0.002628444383541743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,4096,0.014377777775128683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,4096,0.011027555498811932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,768,0.005631999837027655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,3584,0.08280266655815972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,1024,0.005618666609128316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,768,0.003960000144110786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,7168,0.012628444366984896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,32,0.002562666725781229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,65536,0.18485689163208008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,512,0.0034808889031410217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,32,0.0026897777699761917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,3584,0.012974222501118978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,16384,0.07208622164196439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,1536,0.2855191230773926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,8192,0.038422223594453596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,3584,0.02073955535888672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,512,0.011071110765139261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,2560,0.10105511214998032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,7168,0.025193777349260118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,768,0.0047688889834615915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,128,0.005059555586841372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,768,0.016913778252071805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,2560,0.018075555562973022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,1024,0.006130666782458623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,1024,0.04987644486957126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,3072,0.06536355283525255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,8192,0.6473191049363878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,65536,0.18902133570777047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,5120,0.016960889101028442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,1024,0.006044444524579578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,8192,0.01369066701995002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,128,0.006383111079533895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,65536,0.06901688708199395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,512,0.007615110940403408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,2048,0.008019555773999956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,6144,0.012526222401195101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,3584,0.00776444458299213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,3584,0.01348000019788742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,6144,0.0063742221229606206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,512,0.005104889058404499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,2048,0.01238222254647149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,256,0.005233777893914117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,1024,0.011690666278203329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,5120,0.01774577796459198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,768,0.003677333394686381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,1024,0.017673777209387887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,768,0.0037377778854635027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,3072,0.013603555659453074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,6144,0.2596568796369764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,10240,0.7269804212782117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,64,0.00629066675901413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,3584,0.00961599995692571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,32,0.010040000081062317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,5120,0.013117333253224691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,3072,0.009857777920034196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,768,0.003846222327815162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,64,0.0030942221896515954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,512,0.003510222252872255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,1024,0.010503110786279043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,64,0.012840000291665396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,4096,0.010138666464222802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,65536,0.09643289115693833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,2560,0.015747555428081088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,32,0.0052488889131281115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,2048,0.016114667057991028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,12288,0.04353688822852241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,768,0.0047306666771570844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,2560,0.006211555666393704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,1536,0.006217777729034424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,10240,0.013178666432698568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,768,0.07508711020151775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,32,0.005286222116814719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,64,0.005411555369695027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,1024,0.006818666640255187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,65536,0.6889680226643881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,3584,0.03578044308556451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,32,0.008003555238246918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,128,0.0058559998869895935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,256,0.002938666691382726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,512,0.044904887676239014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,1536,0.022190221481853064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,8192,0.07364000214470758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,12288,0.032073779238594904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,3584,0.009330666727489894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,64,0.008585777547624376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,128,0.004001777619123459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,1536,0.008529778156015608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,2560,0.014260444376203748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,32,0.006528000036875407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,65536,0.12900355127122667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,1024,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,128,0.10263911220762466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,10240,1.7666746775309246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,5120,0.025053333905008104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,5120,0.01332355539004008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,65536,0.36820443471272785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,512,0.016549333930015564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,16384,0.030510223574108545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,256,0.0027831111931138565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,5120,0.08879199955198501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,8192,0.03882311118973626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,12288,0.06960711214277479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,4096,0.00830577810605367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,12288,0.02404266595840454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,2560,0.006738666858938005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,3584,0.006843555718660355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,1024,0.004013333469629288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,1536,0.004752000172932942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,6144,0.26934756173027885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,12288,0.9524204466078016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,10240,0.012499555945396423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,32,0.010797333386209277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,10240,0.04794933398564657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,2048,0.00869777798652649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,768,0.004863111095296012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,512,0.004152889053026835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,5120,0.057413332992129855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,512,0.0034640000926123727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,256,0.005524444497293896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,3584,0.01110311100880305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,16384,0.09871644443935818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,512,0.0047564444442590075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,5120,0.011839110818174152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,3584,0.008539555801285638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,8192,0.007346666521496243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,8192,0.014039110806253223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,2560,0.0085235552655326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,768,0.0058248887459437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,256,0.004376888689067629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,10240,0.015541333291265698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,128,0.002733333243264092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,1024,0.007050666544172499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,3072,0.00887288898229599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,32,0.002642666714058982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,5120,0.03924622138341268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,512,0.013366222381591797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,12288,0.032799111472235784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,32,0.0034248888906505373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,10240,0.02857066525353326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,1536,0.10329422023561265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,128,0.004650666481918759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,6144,0.013448889056841532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,8192,0.10922044515609741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,6144,0.03341333402527703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,16384,0.04476622078153822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,3072,0.00829777783817715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,3584,0.008711111214425828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,256,0.004699555536111196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,1024,0.0041351111398802865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,8192,0.2961946593390571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,256,0.002924444360865487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,16384,0.050939556625154286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,12288,0.8177484406365289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,12288,0.020230222079488967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,64,0.003616888903909259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,1024,0.011911999848153857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,768,0.0064328888224230866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,512,0.008730666504965888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,1024,0.007903999752468532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,16384,1.8067012363009984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,128,0.003583999971548716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,768,0.010045333041085137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,7168,0.013052444491121503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,65536,0.30145422617594403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,8192,0.01535288823975457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,6144,2.470370610555013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,128,0.003839111162556542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,4096,0.13217777676052517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,1024,0.0042462220622433555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,64,0.005437333136796951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,4096,0.015035554766654968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,6144,0.013573333621025085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,3072,0.11088266637590195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,8192,0.01713511182202233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,3072,0.01497155593501197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,3584,0.028528889020284016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,1536,0.005055111315515307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,3584,0.007074666519959767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,3072,0.009500444763236577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,10240,0.039170665873421565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,10240,0.026983999543719824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,64,0.008767999708652496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,1024,0.005838222387764189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,7168,0.03870933254559835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,768,0.0034968890249729156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,512,0.005818666683302985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,2048,0.010942222343550788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,1024,0.003923555629120933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,65536,0.0655991103914049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,512,0.006052444378534953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,768,0.003933333274390963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,512,0.004242666479614046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,1536,0.06513422065311007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,256,0.0052622221410274506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,5120,0.16089688407050237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,3584,0.006603555546866522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,12288,0.020094222492641874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,6144,0.012710221939616732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,2048,0.016963554753197562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,256,0.005408888889683618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,128,0.0033617777129014335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,7168,0.01125600023402108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,768,0.05298044284184774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,1536,0.00638755535085996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,256,0.23913065592447916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,3584,0.04733777708477444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,5120,0.23186310132344565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,65536,0.07859200239181519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,5120,0.03495733274353875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,1024,0.010199999643696679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,16384,0.7294409010145398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,1536,0.006768888897365994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,7168,0.028669334120220605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,256,0.006257777826653586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,16384,3.3400647905137806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,3584,0.009137777818573846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,4096,0.017151999804708693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,2560,0.009076444639099969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,16384,0.03372088736957974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,1536,0.00591555568906996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,64,0.00655644428398874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,256,0.023544889357354905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,6144,0.023888889286253188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,5120,0.006600889066855113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,2048,0.007664889097213745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,1024,0.013542222479979197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,32,0.00508977762526936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,3072,0.007607999775144789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,512,0.008225777910815345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,1536,0.011373333632946014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,3072,0.03648266527387831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,6144,0.02334044377009074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,4096,0.02331200076474084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,128,0.0041751112374994485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,65536,0.12240800592634414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,768,0.005600888695981767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,65536,0.010991999672518836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,5120,0.061836441357930504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,3584,0.018668444620238412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,32,0.005229333208666907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,3072,0.016561778055297006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,6144,1.7800159454345703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,2048,0.05156177613470289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,10240,0.037604444556766085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,8192,0.008772444393899705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,768,0.0067528887755341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,6144,0.26756355497572154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,65536,0.29088444179958767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,16384,0.02513244416978624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,10240,0.0350853337181939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,128,0.014597333139843412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,2048,0.007328000333574083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,16384,0.006774222271309958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,512,0.03462933169470893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,8192,0.01706755492422316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,2048,0.0614302224583096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,768,0.057222220632765025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,12288,0.017069333129458956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,6144,0.006325333482689328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,6144,0.006193777753247156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,512,0.015949333707491558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,128,0.005220444252093633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,128,0.0027502222607533136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,64,0.005186666631036335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,512,0.003087999920050303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,1536,0.006824000014199152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,16384,0.0730817781554328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,6144,0.27592266930474174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,8192,0.06662933031717937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,12288,0.38247201177808976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,8192,0.013980444934633044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,512,0.009391110804345872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,2560,0.16615377532111272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,3584,0.01424533294306861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,6144,0.012278222375445895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,7168,0.020764445265134174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,4096,0.00872622181971868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,8192,0.10289688905080159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,6144,0.01016444464524587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,64,0.02001066671477424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,2560,0.021207999851968553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,12288,0.047685331768459745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,7168,0.010480889015727572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,65536,0.06789155801137288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,2048,1.5929235882229273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,64,0.002625777696569761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,256,0.0037546666959921518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,1024,0.025955556167496577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,7168,0.0064088888466358185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,64,0.005330666485759947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,768,0.005816000203291575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,3072,0.11166932847764756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,5120,0.00793866647614373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,256,0.012748444245921241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,1024,0.007222221957312689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,256,0.004808889081080754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,16384,0.10499378045399983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,16384,0.02258488867017958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,3584,1.5342577828301325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,768,0.0058604445722368025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,256,0.006520888871616787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,7168,0.02221422228548262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,1024,0.004403555558787452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,768,0.009217778013812171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,1024,0.023574221465322707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,7168,0.012711111042234631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,10240,0.009172444542249044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,128,0.004578666554556953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,10240,0.017290666699409485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,64,0.008375110725561777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,128,0.0030088888274298776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,3584,0.017960889471901786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,128,0.005976888868543837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,1536,0.01000355515215132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,64,0.0046897778908411665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,128,0.005015111217896144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,7168,0.01886044442653656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,768,0.0069475554757648045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,16384,0.3069244490729438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,3072,0.0074879998962084455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,2048,0.012321777641773224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,8192,0.037815110550986394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,1024,0.005239999956554837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,2048,0.030138668086793687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,5120,0.05778488847944471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,256,0.0028906667398081887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,10240,0.03153866529464722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,768,0.003620444486538569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,7168,0.022837332553333704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,6144,0.01000088867213991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,2560,0.008866666919655269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,4096,0.010737777584128909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,256,0.008393777741326226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,32,0.06338222159279717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,16384,0.01236177815331353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,2560,0.008356444537639618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,65536,1.29566404554579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,4096,0.07971822553210788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,3072,0.00637511122557852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,1536,0.08814222282833523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,12288,0.02208711041344537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,32,0.003087999920050303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,128,0.008897778060701158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,7168,0.6617635620964898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,16384,0.07624889082378812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,12288,0.0169813334941864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,256,0.008804444637563493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,2560,0.009843555589516958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,1536,0.009704888694816166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,3072,0.006339555399285422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,6144,0.040692445304658674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,7168,0.03140799866782294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,64,0.004450666821665234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,5120,0.04596266812748379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,768,0.008565333154466417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,1024,0.011399110986126794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,16384,0.04963199959860908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,65536,0.017708443933063082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,2560,0.02106844385464986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,512,0.0034479999707804788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,256,0.0031102223114834893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,2560,0.0074675555030504865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,5120,0.02481777800454034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,2048,0.005393777870469623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,1536,0.006365333166387346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,6144,0.00867555538813273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,3584,0.006780444333950679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,8192,0.011952888634469775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,65536,0.16856711440616182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,4096,0.012591111163298288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,8192,0.04162133402294583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,64,0.1824746661716037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,1536,0.008783111141787635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,4096,3.3490257263183594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,7168,0.02510133385658264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,10240,0.5614906416998969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,256,0.0480951103899214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,7168,0.010647111468844943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,5120,0.04354666670163473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,10240,0.015576000014940897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,16384,0.006996444529957241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,512,0.004629333400064045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,4096,0.009899555808968013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,10240,0.2814257674747043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,3584,0.023822221491071913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,128,0.0030871110243929755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,1024,0.006095110956165526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,3584,0.019692444139056735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,16384,0.07392444213231404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,12288,1.516930686102973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,512,0.006048000107208888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,1536,0.006084444622198741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,256,0.004292444636424382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,1536,0.006130666782458623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,32,0.0038240001433425476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,16384,0.030113776524861652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,1024,0.013677332964208392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,3584,0.00665511108107037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,10240,0.013088888592190213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,65536,0.31020622783237034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,16384,0.023268444670571223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,32,0.00815822184085846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,8192,0.10768444670571221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,7168,0.018139556050300598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,256,0.008062221937709385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,8192,0.030039999220106337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,768,0.029847999413808186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,256,0.00299466670387321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,5120,0.0381297767162323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,3584,0.014136888914638095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,256,0.005157333281305102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,3584,0.01257333325015174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,128,0.003313777761326896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,2560,0.013917333549923368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,2048,0.06327999962700738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,1536,0.00720711135201984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,2560,0.013400000002649097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,2048,0.004986666556861666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,512,0.007056000332037608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,10240,0.06601422362857394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,7168,0.017416889468828838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,1024,0.009453333086437648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,16384,0.02530577778816223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,1536,0.28735023074679905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,128,0.0034453332838084963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,32,0.009737778041097853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,512,0.9178373548719617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,4096,0.5146737628512913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,8192,0.1533555587132772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,12288,0.4490248892042372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,128,0.003072000005178981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,32,0.002942222274012036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,3072,0.009927111367384592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,1536,0.006718222051858902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,2560,0.007165333463086023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,5120,0.017180444465743173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,128,0.004610666798220741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,128,0.006159111029571957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,1024,0.006778666542636023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,12288,0.03288177649180094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,2048,0.008436444732877944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,128,0.005499555418888728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,1024,0.027599111199378967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,1536,0.006051555689838197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,5120,0.009925333162148794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,768,0.015917332635985482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,12288,0.015223999818166098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,5120,0.02161688937081231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,12288,0.1445653306113349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,3584,0.024096889628304377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,8192,0.1743884483973185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,32,0.004632888982693355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,256,0.004401777767472797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,64,0.002504000026318762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,12288,0.15081777837541369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,16384,0.026833777626355488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,8192,0.015871110889646742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,10240,0.06889688968658447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,4096,0.015227556228637695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,768,0.004229333251714706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,128,0.00434933313065105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,2560,0.005975111077229182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,3072,0.009320889082219865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,768,0.01573066744539473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,6144,0.024430221981472437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,2048,0.010460444622569613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,512,0.0073724447025193115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,12288,0.0058044443527857465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,256,0.007520000139872233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,65536,0.21790577305687797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,32,0.002653333254986339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,4096,0.013528889252079858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,10240,0.01604711181587643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,2048,0.02375999920898014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,64,0.006430222342411677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,1536,0.011158222125636207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,2048,0.013761777844693927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,512,0.017909333109855652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,6144,0.01016533292002148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,64,0.005676444619894028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,8192,0.011424889167149862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,12288,0.05288622114393446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,2560,0.007592000067234039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,2560,0.02056711084312863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,12288,0.016414221790101793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,768,0.006592000110281839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,7168,0.023144889209005568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,64,0.005927111125654644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,3584,0.016742222838931613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,2048,0.01920355525281694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,6144,0.025060445070266724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,12288,0.0929866698053148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,65536,0.029494222667482164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,128,0.003991111285156674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,4096,0.012495111260149213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,768,0.005548444473081165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,65536,0.05886399745941162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,5120,0.03530755639076233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,7168,0.014529777897728814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,8192,0.006640000061856375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,12288,0.01680177781316969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,16384,0.006274666637182236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,256,0.015790222419632804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,8192,0.012887111140622033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,10240,0.06108355522155762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,10240,0.013271110753218332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,256,0.008612444831265343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,7168,0.019948444432682462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,2048,0.07042044401168823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,2560,0.009392889009581672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,768,0.0058142224119769196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,1024,0.0058959999846087555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,16384,0.00871377769443724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,8192,3.4804195827907987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,2560,0.030068443881140813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,768,0.007890666524569193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,6144,0.015632889337009855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,12288,0.18003555138905844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,256,0.003656889001528422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,64,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,8192,0.35097511609395343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,2560,0.013011555704805585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,8192,0.22704890039232042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,128,0.007646222081449296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,65536,0.2858560085296631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,6144,0.03559022148450216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,6144,0.01223111069864697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,1024,0.0063982220987478895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,3072,0.08609600199593438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,32,0.006781333436568578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,512,0.010090666512648264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,5120,0.008752889103359645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,6144,0.006040888941950268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,768,0.00527022240890397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,1536,0.006414222220579783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,65536,0.0728435582584805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,1536,0.0063457778758472866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,3072,0.007684444387753804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,10240,0.041286223464541964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,256,0.005097777893145879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,768,0.003608889049953885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,10240,0.020676443974177044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,2560,0.02576622201336755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,768,0.018668444620238412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,4096,0.008389333056079017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,1536,0.009034666750166152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,5120,0.009335999687512716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,12288,0.044109334548314415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,256,0.0029928889125585556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,16384,0.019665777683258057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,128,0.011064888702498542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,12288,0.040807998842663236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,1536,0.006778666542636023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,5120,0.018556444181336295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,64,0.007786666353543599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,5120,8.251095241970486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,512,0.005893333504597346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,64,0.008591999610265097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,1536,0.00572622236278322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,16384,0.05564888980653551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,7168,0.3971155484517415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,768,0.004658666749795278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,5120,0.028951111767027114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,2048,0.005240889059172736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,10240,0.02892177634769016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,6144,0.010464888479974536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,12288,0.00925688859489229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,10240,0.02157955533928341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,768,0.01017955525053872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,8192,0.030026667647891577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,128,0.004922666483455234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,4096,0.006055999961164262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,768,0.010220444036854638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,65536,0.41849599944220645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,2560,0.04215555720859104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,65536,2.199839062160916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,128,0.005398222141795688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,3072,0.03347911106215583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,8192,0.011411555939250521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,8192,0.24692355261908638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,512,0.004650666481918759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,4096,0.014347554908858405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,1024,0.04968355430497063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,512,0.00562755556570159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,12288,0.07082311312357585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,12288,0.04395111070738899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,3072,0.010963555839326648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,2560,0.19859822591145834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,12288,0.05343022280269199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,3584,0.01627111103799608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,8192,0.011898666620254517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,1536,0.014158222410413953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,512,0.003608889049953885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,4096,0.0278479986720615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,65536,0.3320506678687202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,4096,0.011134222149848938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,2560,0.6864186922709147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,2048,0.008623110751310984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,1024,0.006800888727108638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,512,0.003436444534195794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,3584,0.018802666001849704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,65536,0.1553057829538981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,64,0.008760000268618265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,8192,0.014164444473054675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,1536,0.00675733346078131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,12288,0.015231110983424716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,16384,0.11296889517042373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,8192,0.06877688566843669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,10240,0.01632355484697554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,8192,0.012092444631788464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,512,0.01069866700304879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,1024,0.1538755496342977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,768,0.005842666659090254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,5120,0.01216533366176817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,2048,0.317392004860772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,2048,0.009335111412737105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,512,0.007807999849319458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,128,0.10143911176257664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,5120,0.008671110702885522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,3072,0.2421466774410672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,32,0.002556444456179937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,8192,3.4492986467149525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,256,0.009741333623727163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,6144,0.018795554836591084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,16384,0.022430222895410325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,1536,0.06767911381191678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,2560,0.009229333036475712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,16384,0.28175822893778485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,512,0.030173334810468886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,512,0.0047395556337303585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,7168,0.014452444182501899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,4096,0.017628444565667045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,3584,0.008880889250172509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,7168,0.019430221782790292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,65536,0.14628977245754665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,256,0.012727110750145383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,768,0.0034302222645945023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,64,0.003269333392381668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,7168,0.04797066582573784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,8192,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,12288,0.33712445365058047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,256,0.0028648889727062653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,1024,0.11367644204033746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,512,0.007644444704055786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,1536,0.006125333408514659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,1536,0.006474666711356904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,65536,0.06664088699552748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,10240,0.31612356503804523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,512,0.0074346669846110875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,768,0.005072888814740711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,12288,0.20582755406697592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,10240,0.04515377680460612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,512,0.005798222290145026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,16384,0.03637422124544779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,32,0.002647111192345619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,1024,0.00600000015563435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,7168,0.026567111412684124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,16384,0.20982221762339273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,128,0.02667733364635044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,4096,0.015185778339703878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,3584,0.023927999867333308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,256,0.004606222112973531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,64,0.006365333166387346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,10240,0.019446222318543326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,6144,0.15073600080278185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,8192,0.01865866614712609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,10240,0.017118222183651395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,7168,0.010739555789364709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,6144,0.012513777448071374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,3072,0.008255110846625434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,1024,0.004284444368547863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,2048,0.021125333176718816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,3072,0.02387911081314087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,10240,0.05795377492904663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,2048,0.010141332944234213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,7168,0.013969777358902825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,6144,0.012972444295883179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,12288,0.018021333548757765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,10240,0.010658666491508484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,768,0.007663999994595845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,1024,0.00866844422287411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,2560,0.007669332954618666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,768,0.004846222284767362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,2048,0.0136817776494556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,12288,0.026476444469557867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,32,0.004935111022657818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,65536,0.08254844612545438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,64,0.005063111169470681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,128,0.009393778112199571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,3072,0.12333867284986709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,10240,0.022395556171735127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,5120,0.006588444527652528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,4096,0.009778666827413771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,1024,0.006626666833957036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,3584,0.009350222018029954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,2560,0.010323555933104621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,256,0.01405422223938836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,16384,0.14191999700334337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,6144,0.013495999905798169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,512,0.012552888856993781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,256,0.01776622235774994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,10240,0.03991555505328708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,12288,0.01661422186427646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,128,0.004609777695602841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,8192,0.030333333545260962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,16384,0.008031999899281396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,2048,0.018444443742434185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,128,0.005399999933110342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,2560,0.014174222946166992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,4096,0.08236710892783271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,128,0.009491555392742157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,1536,0.018943111101786297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,16384,0.05619733201132881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,2048,0.019949333535300363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,65536,0.1260319948196411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,768,0.0058000000814596815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,7168,0.03590755661328634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,2560,0.01682933337158627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,128,0.006102222121424145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,1024,0.005884444548024072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,256,0.006777777853939269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,128,0.0027262222849660446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,256,0.003032888803217146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,1536,0.006896889044178857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,4096,0.009898666706350114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,128,0.01569333341386583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,1536,0.0058551111982928375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,6144,0.05904533465703329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,128,0.0041146667467223275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,2560,0.006767999794748094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,6144,0.011648000114493899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,1024,0.009953777823183272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,768,0.005584888988071018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,2560,0.007818666597207388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,65536,0.09300355778800116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,4096,0.04522044459978739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,128,0.006683555742104848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,3072,0.01800622211562263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,8192,0.026943110757403906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,512,0.006276444428496891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,16384,0.03867644402715895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,7168,0.01536977787812551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,1024,0.004594666676388847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,6144,0.1920320060518053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,512,0.0035013332962989807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,768,0.010695999695195092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,7168,0.014357333381970724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,4096,0.008157333566082848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,10240,0.06588177548514472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,6144,0.026181333594852026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,12288,0.015217777755525378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,2560,0.007802666889296637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,2560,0.005977777971161737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,32,0.0052133335007561584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,16384,0.05516089002291361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,10240,0.00849599970711602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,1024,0.0059279998143513995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,768,0.005910222315125995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,2048,0.016200888488027785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,6144,0.013080889152155982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,512,0.003661333272854487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,256,0.005432888865470886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,64,0.0026844443960322272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,32,0.00518133325709237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,4096,0.015415110521846347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,1024,0.010189333723651038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,1024,0.005680888891220093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,5120,0.3384115431043837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,1536,0.006511111226346757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,512,0.01816177699300978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,32,0.0041893331540955436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,32,0.005745777653323279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,3072,0.12643999523586696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,16384,0.14157511128319636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,768,0.011295110815101199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,12288,0.03532800078392029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,10240,0.014152889450391134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,2560,0.035750223530663386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,256,0.006813333266311222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,64,0.0026666666898462507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,32,0.006606222026877933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,256,0.021180444293551978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,768,0.006151111175616582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,32,0.006500444478458828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,10240,0.02142133315404256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,768,0.04483288857671949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,65536,0.5495431158277724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,12288,0.049908445941077344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,8192,0.018965333700180054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,512,0.0075431110130416016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,16384,0.02296977738539378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,3584,0.040510220660103693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,65536,0.07476355632146199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,12288,0.02090577781200409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,5120,0.08611022101508246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,512,0.006884444504976273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,1536,0.0070977773931291364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,256,0.0034044444974925783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,7168,0.013023999830087027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,768,0.008761777646011775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,512,0.0047439999050564235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,1536,0.006773333168692059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,8192,0.1938239998287625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,128,0.054818666643566556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,10240,0.01937333411640591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,4096,0.011876444849703046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,768,0.003693333516518275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,64,0.003530666646030214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,3584,0.022163555026054382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,8192,0.10833955473370022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,768,0.008110221889283922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,5120,0.12961333327823216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,768,0.00452888881166776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,7168,0.05515377720197042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,256,0.005192888693677055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,65536,0.20985599358876547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,6144,0.012415111064910889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,3072,0.009141333401203156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,12288,0.07767022318310209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,1536,0.00926577796538671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,768,0.004567999806669024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,65536,0.07474399937523736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,256,0.0047217777205838096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,12288,0.017867555220921833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,2048,0.013304000099500021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,128,0.002995555599530538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,5120,0.017217778497272067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,512,0.003698666476541095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,512,0.03856089048915439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,768,0.005640000104904175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,1024,0.006688888702127669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,16384,0.10895378059811062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,5120,0.006158222340875202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,10240,0.08439911074108547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,2048,0.005274666680230035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,10240,0.01677688956260681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,64,0.0028613333900769553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,16384,0.05538755655288696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,1536,0.013933334085676404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,16384,0.027540443672074214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,1024,0.007092444433106317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,8192,0.5383129119873047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,12288,0.021314667330847845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,5120,0.022132444712850783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,4096,0.1557448837492201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,2048,0.004936888813972473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,6144,0.008839999636014303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,10240,0.019263111882739596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,2560,0.008965333302815756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,1024,0.003915555361244413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,64,0.0025600000388092464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,2048,0.018200000127156574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,1024,0.01051111105415556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,8192,0.010120000276300643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,768,0.012561777399645912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,1024,0.011366222467687396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,7168,0.20127199755774605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,65536,0.04795733425352308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,10240,0.6315600077311198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,512,0.006414222220579783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,768,0.005685333162546158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,16384,0.032953778902689614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,768,0.006180444525347815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,7168,0.021125333176718816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,1024,0.00612088872326745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,1536,0.0047102222839991255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,128,0.005404444618357553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,64,0.0031555555760860443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,1536,0.010780444575680627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,7168,0.15414844618903265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,6144,0.012772444221708508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,10240,0.016196444630622864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,2560,0.014914666612943014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,2560,0.006575111299753189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,1024,0.014883556299739413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,2048,0.009791110952695211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,5120,0.019840000404251944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,12288,0.01514311134815216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,3072,0.14304089546203613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,256,0.00797155582242542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,64,0.002552888873550627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,16384,0.027098667290475633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,1536,0.043007999658584595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,3584,0.2515244483947754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,7168,0.014507555299335055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,65536,0.08500266737408108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,65536,0.18657067086961535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,512,0.012938666674825879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,65536,0.8755102157592773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,12288,0.049550222025977246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,768,0.05084977878464592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,6144,0.01238222254647149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,6144,0.020187555087937247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,1024,0.00924355536699295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,2560,0.006783111227883234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,6144,0.03294488787651062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,16384,0.08598755465613471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,12288,0.018971555762820773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,7168,0.006242666807439592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,1024,0.005517333332035277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,12288,0.031130668189790513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,512,0.005933333188295364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,8192,0.11200443903605144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,512,0.027229333917299908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,65536,4.082522498236762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,10240,9.519006517198351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,3584,0.007784888976150089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,8192,0.027683556079864502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,768,0.004975999808973736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,4096,0.015805333852767944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,10240,0.07106933328840467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,5120,0.021900445222854614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,12288,0.025456888808144465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,6144,0.03656889001528422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,128,0.017180444465743173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,5120,0.014489778214030795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,512,0.004334222111437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,65536,0.06044977903366089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,3072,0.055191109577814736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,768,0.003945777813593547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,768,0.0058319999112023254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,65536,0.4657413164774577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,3584,0.01148000028398302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,12288,0.023758222659428913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,512,0.007244444555706448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,10240,0.04599022203021579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,65536,1.1490329106648762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,12288,0.052355554368760854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,512,0.01052177780204349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,16384,0.035440001222822405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,3072,0.010149333212110732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,16384,0.020850666695170935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,2048,0.037648889753553606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,256,0.011135111252466837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,65536,0.2912693288591173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,5120,0.0069466663731469055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,10240,0.021606221795082092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,32,0.007078222102589077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,1024,0.009538667069541084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,7168,0.01367022262679206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,5120,0.22029511133829752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,2560,0.011392888923486074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,1536,0.00480266660451889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,16384,0.049264887968699135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,1536,0.15118222766452366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,64,0.002720000015364753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,256,0.005587555468082428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,8192,0.0805964469909668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,3584,0.016719111137919955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,16384,0.04424711068471273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,16384,0.2757839891645643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,16384,0.220360000928243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,128,0.0041537777417235905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,6144,0.02204799983236525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,7168,0.01536177761024899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,4096,0.019867555962668527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,256,0.005435555759403441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,4096,0.036453333165910505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,12288,0.4421413209703233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,256,0.00563466673096021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,64,0.005842666659090254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,1536,0.01834844383928511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,1024,0.007276444799370236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,10240,0.008635555704434713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,4096,0.008858666651778752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,3072,0.04445866743723551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,10240,0.020750222934616935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,1024,0.003998222036494149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,2560,0.009478222164842818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,2048,0.07232622305552165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,1024,0.00963644435008367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,2560,0.015322667029168872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,2048,0.0057911111248864075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,256,0.01094044413831499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,256,0.005507555686765247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,6144,0.01182666669289271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,5120,0.9827191034952799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,4096,0.0571848882569207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,768,0.005559111220969095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,65536,0.1075635552406311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,12288,0.06526133086946276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,2048,0.00865955568022198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,2560,0.01038044442733129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,256,0.0038497779104444715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,512,0.0038204445607132385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,1536,0.022069333328141108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,16384,0.12043555577596028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,8192,1.2465457916259766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,6144,0.017432888348897297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,1024,0.006639110959238476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,4096,0.010593777729405297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,6144,0.02848177817132738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,6144,0.013870221873124441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,5120,0.019719999697473314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,4096,0.010813333094120026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,2560,0.05031822125116984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,65536,1.458939552307129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,256,0.009390222529570261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,5120,0.011262222296661802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,256,0.005782222168313132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,10240,0.07015555434756808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,10240,0.9800702200995551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,1024,0.00996800015370051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,512,0.009530666801664565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,7168,0.024515555964575872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,1024,0.0373911095990075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,10240,0.2580995559692383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,65536,0.5566142400105795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,65536,0.018202667435010273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,256,0.006768888897365994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,3072,0.015105777316623263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,128,0.0031911111954185697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,1024,0.007469333708286285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,256,0.006771555377377405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,512,0.009962666365835402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,7168,0.011795555551846823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,5120,0.019506666395399306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,16384,0.09867288668950398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,10240,0.019204444355434842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,32,0.010285333626800114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,8192,0.0336444444126553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,64,0.006568000134494569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,8192,0.006160000132189856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,2048,0.007495111061467065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,5120,0.006296000132958095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,3072,0.01872711049185859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,4096,0.018554665976100497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,3072,0.013616888887352414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,8192,0.02735200027624766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,3584,0.025241777300834656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,6144,0.02498400045765771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,512,0.007165333463086023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,16384,0.02864177690611945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,5120,0.009581333233250512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,3584,0.1367866728040907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,1536,0.008560889297061497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,768,0.0057964444988303715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,128,0.0028151110228565005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,16384,0.08699644274181789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,65536,0.09665511051813762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,2048,0.024731554918819006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,256,0.02918755677011278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,2560,0.006918222126033571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,65536,0.06522578001022339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,16384,0.11046222183439468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,65536,0.18678400251600477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,2560,0.014316444595654806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,16384,0.08203288581636217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,10240,0.020748444729381137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,8192,0.01901244454913669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,2048,0.7538880242241753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,768,0.00590133335855272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,6144,0.19308355119493273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,2048,0.009407111340098912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,128,0.008183110919263627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,256,0.03090311090151469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,4096,0.010268444816271463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,64,0.005383111122581694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,1024,0.005745777653323279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,4096,0.008329778081840938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,3584,0.008185777399275038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,3584,0.023221333821614582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,768,0.004979555391603046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,32,0.0035626664757728577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,65536,0.08965422047509088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,512,0.006143110907740063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,768,0.005852444304360284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,2048,0.012524444195959302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,5120,0.013624889155228933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,4096,0.01881866653760274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,2048,0.02032088902261522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,1536,0.009730666875839233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,6144,0.12025066216786702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,4096,0.006248888870080312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,512,0.004835555536879433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,3072,0.012208888928095499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,4096,0.011311111350854238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,1536,0.004994666824738185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,12288,0.015166223049163818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,32,0.012152888708644442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,1024,0.01000088867213991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,128,0.0058471109304163195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,8192,0.012695999609099494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,3072,0.02077511118517982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,3584,0.26550134023030597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,4096,0.008113777471913232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,8192,0.029350221157073975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,12288,0.02674399978584713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,64,0.02223377757602268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,32,0.0063902222447925145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,64,0.007082666787836287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,6144,0.03130311104986403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,512,0.004384888956944148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,12288,0.026744888888465032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,4096,0.01297155519326528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,1536,0.009533333281675974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,12288,0.015378667248619927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,3072,0.0881706674893697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,1024,0.009583111438486312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,7168,0.046446220742331616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,32,0.005655999812814925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,4096,0.018086221483018663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,768,0.01094133324093289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,2048,0.391682677798801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,7168,0.029008888536029395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,128,0.003638222399685118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,8192,0.04039288891686334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,6144,0.020483555065260995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,5120,0.027469333675172593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,10240,0.018912000788582694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,3584,0.10347200102276272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,2560,0.20309421751234266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,2048,0.03294310967127482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,2560,0.006464889066086874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,10240,0.027199111051029626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,12288,0.03171022070778741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,65536,0.06691822078492907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,3584,0.015070221490330167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,12288,0.0763182242711385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,3072,0.013070222404268054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,5120,0.017685333887736004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,6144,0.010584889186753167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,1536,1.3670916027492945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,2048,0.011143111520343356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,10240,0.09712888797124226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,128,0.003589333345492681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,512,0.0068604445291890045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,4096,0.007986666427718269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,32,0.003108444520168834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,5120,0.09624444113837348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,1536,0.011591999895042844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,2048,0.016460445192125108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,64,0.006517333288987477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,2560,0.018742221924993727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,1024,0.005975111077229182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,10240,0.10179911057154338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,7168,0.016545777519543965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,10240,0.036121779017978244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,3584,0.03534933262401157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,16384,0.018557333283954196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,12288,0.2650044494205051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,10240,0.22037955125172934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,768,0.005769777629110549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,3072,0.012293332980738746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,3584,0.015371556083361307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,2048,0.013656888571050433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,8192,0.37932711177402073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,12288,0.006440889090299606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,3072,0.02036977807680766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,6144,0.011149333582984077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,512,0.004990222139490975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,5120,0.10198844141430324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,8192,0.02316355539692773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,2560,0.018618666463428073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,2560,0.007460444337791867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,6144,0.011022222538789114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,2048,0.006093333164850871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,512,0.006068444500366847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,3072,0.00647199981742435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,2560,0.10681688785552979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,4096,0.22511288854810926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,7168,0.041173332267337374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,512,0.03376266691419814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,4096,0.025831111603313025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,5120,0.0068417779273457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,128,0.002776888923512565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,2048,0.013344888885815939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,2560,0.007914666500356462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,1536,0.038719111018710665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,65536,0.14023378160264757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,3072,0.014937778313954672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,1024,0.004853333450025982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,512,0.005686222265164058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,128,0.0025439999169773525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,128,0.003342222215400802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,256,0.007196444604131911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,1536,0.009662222531106737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,512,0.004362666772471534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,4096,0.00920800036854214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,64,0.008859555754396651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,768,0.12288533316718207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,6144,0.04623022344377306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,128,0.0040693332751592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,65536,10.880171881781685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,4096,0.03301422132386102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,32,0.0034151110384199354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,3072,0.014782221780882942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,12288,0.10812355412377252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,3072,0.0070088886552386815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,16384,0.02714666724205017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,128,0.004425777743260066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,10240,0.026339555780092876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,128,0.005655999812814925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,1024,0.004612444589535396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,4096,0.006299555715587404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,6144,0.1039724416202969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,7168,0.02730577852990892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,7168,0.015262222952312894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,256,0.005559999909665849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,3584,0.01758666667673323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,128,0.006026666611433029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,16384,0.02049422264099121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,256,0.002895111011134254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,4096,0.28965777821011013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,4096,0.014001778430408902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,7168,0.01905333333545261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,768,0.035275555319256253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,256,0.03934933410750495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,3072,0.05062311225467258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,10240,0.018834667073355783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,6144,4.949576059977214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,5120,0.01056177748574151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,128,0.006140444427728653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,128,0.05280177791913351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,16384,0.006610666712125142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,1024,0.026050666968027752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,2560,0.012166221936543783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,2048,0.014814222852389017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,4096,0.010467555787828235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,5120,0.015240000353919135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,7168,0.01462044484085507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,4096,0.01052266690466139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,1024,0.15530044502682155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,512,0.004941333499219683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,512,0.005250666704442766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,64,0.0023973333752817577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,512,0.005665777872006099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,256,0.006292444550328785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,2560,0.01831288966867659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,256,0.005365333209435145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,3584,0.01607644392384423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,128,0.004015111260943943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,16384,0.027635556128289964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,8192,0.015593777100245157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,512,0.005072000126043956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,2560,0.014519110321998596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,2560,0.05814933114581638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,65536,0.5904071066114638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,128,0.007033777733643849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,10240,0.028799109988742407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,1536,0.014697778556081982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,128,0.004990222139490975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,3584,0.006280000011126201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,2048,0.005214222189452913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,12288,0.04456355505519443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,768,0.009489778015348647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,3072,0.014907555447684394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,7168,0.025359110699759588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,7168,0.1461271047592163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,256,0.008631111019187504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,2048,0.007490666376219855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,10240,0.02792444494035509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,1536,0.005454222361246745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,4096,0.01014755583471722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,1536,0.028605331977208454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,65536,0.07748355468114217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,6144,0.013395555317401886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,1536,0.007738666401969061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,65536,0.04297866755061679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,3072,0.03730577892727322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,128,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,3584,0.012467555701732635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,12288,0.037662221325768366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,2048,0.012645333177513547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,12288,0.03327200147840712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,4096,0.009923555784755284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,5120,0.01402933398882548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,1024,0.003948444293604957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,8192,0.019336000084877014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,128,0.007739555504586961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,6144,0.014836443795098199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,512,0.003678222083383136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,1024,0.006172444257471297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,8192,0.013516444298956128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,7168,0.012614222036467658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,2560,0.12688533465067545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,7168,0.01127822200457255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,6144,0.02104622291194068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,16384,0.007835555407736037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,768,0.00870488915178511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,8192,0.015810666812790763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,3072,0.02325155503220028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,3584,0.00756177802880605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,65536,0.563477357228597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,6144,0.23621156480577257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,2560,0.09051555395126343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,256,0.005618666609128316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,512,0.004566222015354368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,256,0.005772444523043103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,8192,0.01905955539809333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,8192,0.043860445419947304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,5120,0.0317537784576416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,2560,0.005499555418888728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,12288,0.06136622031529745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,768,0.0034533333447244433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,6144,0.023895111348893907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,5120,0.006949333681000604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,10240,0.04408266809251574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,16384,0.5617146492004395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,768,0.006508444332414203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,1024,0.005165333135260476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,8192,0.06555822160508898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,256,0.004472889006137848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,2048,0.049117333359188504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,6144,0.01570133368174235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,8192,0.008044444852405125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,256,0.005187555733654234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,12288,0.5969626638624403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,768,0.008051555189821456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,768,0.006775110960006714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,1024,0.22282756699456108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,2560,0.02151288919978672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,3072,0.010309333602587381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,12288,0.021962665849261816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,512,0.003979555434650845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,65536,0.50436798731486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,3584,0.008049777812427944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,768,0.0074142225914531285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,6144,0.05448089043299357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,768,0.008403555386596257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,8192,0.017469333277808297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,2048,0.23259109920925566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,5120,0.01685333251953125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,10240,0.04780888888570997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,768,0.008302222523424361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,1536,0.014762666490342883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,128,0.00434755575325754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,65536,0.40855201085408527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,512,0.0046977777447965415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,128,0.005872888697518243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,7168,0.2854222191704644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,3072,0.08748800224728054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,2048,0.018008889423476327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,10240,0.13421600394778782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,3584,0.01900800069173177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,12288,0.02102577851878272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,1024,0.003949333396222857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,4096,0.008749333520730337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,3072,0.021398221453030903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,64,0.004943111290534337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,32,0.005092444519201915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,128,0.006485333459244833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,16384,0.023403555154800415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,3584,0.007907555335097844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,3584,0.11540444691975911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,512,0.005625777774386936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,32,0.005134222408135732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,4096,0.012431110772821637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,6144,0.009920888476901585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,12288,0.03202666507826911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,7168,0.01220355596807268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,32,0.006601777755551868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,65536,0.08795911073684692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,512,0.006720888945791457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,7168,0.028355555401908025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,3584,0.014581332604090372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,4096,0.03187911046875848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,768,0.005746666755941179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,3072,0.015207999282413058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,65536,0.06958666774961683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,5120,0.007495111061467065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,3584,0.013433777623706393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,10240,0.08004888561036852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,65536,0.2566693358951145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,5120,0.17584799395667183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,8192,0.006702222343948152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,6144,0.05830933650334676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,16384,1.2922675874498155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,2560,0.11834311485290527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,10240,0.01330044451687071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,7168,0.21875554985470244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,8192,0.012106666962305704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,64,0.002942222274012036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,3584,0.013751111096805997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,3584,0.009863999982674917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,2560,0.02769244379467434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,4096,0.02311733365058899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,3072,0.0151235560576121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,64,0.006663110935025745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,2560,0.015237333046065437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,128,0.005473777651786804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,8192,0.09288622273339166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,2560,0.013293333351612091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,10240,0.024093333217832778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,512,0.005315555466545953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,12288,0.02184533410602146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,5120,0.012340444657537671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,65536,0.3134862316979302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,768,0.010247111320495605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,6144,0.006255110932721033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,5120,0.09765333599514431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,5120,0.021576000584496394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,768,0.005598222215970357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,4096,0.01658933361371358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,128,0.008201777935028076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,128,0.02474133339193132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,16384,0.5082409116956923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,16384,0.5161271095275879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,64,0.005607999861240387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,2560,0.009156444834338294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,6144,0.39946667353312176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,128,0.004823110997676849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,768,0.007656888829337225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,8192,0.019075555933846366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,6144,0.05697333150439792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,8192,0.2772382100423177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,3584,0.009742221898502773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,768,0.022612444228596155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,3584,0.007328000333574083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,1024,0.006118222243256039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,65536,0.07730666796366374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,1536,0.011161777708265515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,1536,0.009915555516878763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,128,0.0028613333900769553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,768,0.03340977761480544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,3584,0.008912000391218398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,8192,0.15968888335757786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,65536,0.18191288577185738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,2560,0.038923554950290255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,2560,0.005970666805903117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,10240,0.039572444227006696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,12288,0.017861333158281114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,8192,0.0562115576532152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,8192,0.02308888898955451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,65536,0.19299466080135771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,7168,0.0410462220509847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,32,0.004646222210592694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,5120,0.02241866621706221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,6144,0.013286222186353473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,65536,0.07760710848702325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,8192,0.13538400332132974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,512,0.005541333307822545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,2048,0.0075662218862109715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,2560,0.005931555396980709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,12288,0.04494222336345249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,12288,0.0151591118839052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,512,0.006826666494210561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,10240,0.06587111287646823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,4096,0.011649777491887411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,4096,0.009977777798970541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,1536,0.00720000018676122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,1536,0.006215999937719769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,6144,0.06711377700169881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,2560,0.17115999592675102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,768,0.02126666737927331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,768,0.023353776997990076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,3584,0.06239289045333862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,3072,0.013923555612564087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,512,0.00563733321097162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,6144,0.009535111486911774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,16384,0.022776888476477727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,512,0.009894222021102905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,1536,0.004625777817434735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,32,0.003097777772280905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,128,0.014463999205165438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,4096,0.015986666083335876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,16384,0.03372800019052293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,3072,0.014343111051453484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,12288,0.05728533532884386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,10240,0.040525333748923406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,1024,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,16384,0.2103644476996528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,3584,0.009408888717492422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,32,0.00304888892504904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,1024,0.005081777771313985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,10240,0.0655262205335829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,3072,0.030682666434182063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,512,0.003664888855483797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,1536,0.010837333069907295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,768,0.005459555735190709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,1536,0.0046826667255825466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,2560,0.005580444302823808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,128,0.0053688887920644544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,16384,26.609490288628475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,512,0.020601777566803824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,6144,0.014700444208251106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,2048,0.006516444600290722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,7168,0.02216533323129018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,512,0.01677600045998891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,512,0.008927110996511247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,12288,0.017180444465743173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,2560,0.007171555525726742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,128,0.005112888912359874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,2560,0.006036444256703059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,5120,0.01200799975130293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,65536,0.13775732782151964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,12288,0.006214222146405115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,512,0.004197333421972063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,3584,0.17874044842190215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,16384,0.14401155047946507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,4096,0.15458222230275473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,5120,0.07002844413121541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,2048,0.007941332956155142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,4096,0.025191111697090998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,3584,0.008551999926567078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,768,0.0038053331275780997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,128,0.004285333471165763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,12288,0.05872978104485405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,3072,0.04195644458134969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,32,0.0026506665680143568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,12288,0.00981422182586458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,12288,0.027101332942644756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,128,0.003994666867785984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,7168,0.012337778177526263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,4096,0.031160000297758315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,2560,0.01610488858487871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,256,0.008296888735559251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,12288,1.03657775455051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,768,0.005173333403136996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,5120,0.010371555884679159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,8192,0.6925324334038628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,10240,0.059522668520609535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,12288,0.007892444729804993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,12288,0.41120089424981016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,10240,0.12515555487738714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,3072,0.006153777655627992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,3584,0.017447999782032438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,3072,0.05420355664359199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,128,0.05017955435646904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,2560,0.005903111149867375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,64,0.013834666874673633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,12288,0.02144177754720052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,6144,0.02940533227390713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,1024,0.004063111212518481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,256,0.0032995556377702286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,768,0.0052435555391841466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,32,0.007204444044166141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,12288,0.02912799848450555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,65536,0.05923644701639811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,3072,0.008003555238246918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,768,0.005470222069157495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,2560,0.007541333635648091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,768,0.005583111196756363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,512,0.04709600077735054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,12288,0.15385599931081137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,2560,0.007721777591440413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,5120,0.012914666699038612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,1536,0.006311999840868845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,10240,0.027615999182065327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,256,0.003043555551105075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,7168,0.011200000014570026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,768,0.004253333227501975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,1024,0.03151999910672506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,256,0.005404444618357553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,5120,0.4470444255405002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,256,0.010187555518415239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,2560,0.00793422261873881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,32,0.007489778101444244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,5120,0.014492443866199918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,3072,0.012810666528013019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,2048,0.04233688778347439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,10240,0.013156444662147097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,128,0.007651555869314406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,6144,0.009278222090668148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,3072,0.018290667070282828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,64,0.0027155555370781156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,128,0.006514666808976068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,2048,0.012658666405412885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,10240,0.7623457908630371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,4096,0.008767111433876885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,16384,0.043212443590164185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,8192,0.15988266468048096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,12288,0.2642729017469618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,128,0.0047146665553251905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,65536,0.07951822545793322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,8192,0.03161244591077169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,3072,0.2400968869527181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,7168,0.03447733322779337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,2560,0.06740178002251519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,3072,0.01774044500456916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,5120,0.006952889263629913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,12288,0.0326257778538598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,7168,0.0292595558696323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,2560,0.0087333329849773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,65536,0.13617422845628527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,768,0.006313777632183499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,768,0.004842666702138053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,12288,0.40809154510498047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,5120,0.015293333265516492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,1536,0.005821333163314396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,65536,0.14817244476742214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,65536,0.08018577761120267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,3584,0.008206222620275285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,6144,0.014167999227841696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,65536,0.1846595605214437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,64,0.005868444426192178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,12288,0.017509332961506315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,512,0.003666666646798452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,4096,0.007728888756699032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,1024,0.012432888978057437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,2048,0.007798222204049428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,6144,0.007673777639865875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,1536,0.0046773333516385825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,1536,0.011811555259757571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,65536,0.2805573410458035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,16384,0.0525511105855306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,128,0.004121777911980947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,7168,0.006055999961164262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,5120,0.15970310899946424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,7168,0.01443555619981554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,32,0.022286223040686712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,4096,0.012282667060693106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,1536,0.010865777730941772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,4096,0.022423999177085027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,2560,0.05814844369888306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,65536,0.077674667040507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,5120,0.009158222211731805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,10240,0.0655848913722568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,256,0.0053546664615472155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,512,0.003540444291300244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,5120,0.006761777732107375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,2560,0.0075697774688402815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,5120,0.010159999959998662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,16384,0.025778666138648987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,64,0.004506666627195147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,7168,0.01408622165520986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,12288,0.0395742224322425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,6144,0.0176862229903539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,12288,0.03100799851947361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,12288,0.0683057771788703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,128,0.0034551111360390982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,16384,0.06050933069652981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,6144,0.0138586668504609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,65536,0.2944657802581787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,7168,0.007095110913117726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,10240,0.006107555495368109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,3584,0.03302044338650174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,3072,0.03180888957447476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,7168,0.14859378337860107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,768,0.00609422226746877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,8192,0.04983289043108622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,1536,0.00462755560874939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,4096,0.008349333372380998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,768,0.014955555399258932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,65536,0.20608088705274794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,1024,0.006044444524579578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,10240,0.3738364378611247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,5120,0.018631999691327412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,2560,0.007238222493065729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,768,0.005992888990375731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,256,0.003730666720204883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,6144,0.01644888851377699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,2048,0.005869333528810077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,3584,0.010797333386209277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,65536,0.2559537887573242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,3072,0.012436444560686747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,3584,0.06208178069856432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,8192,0.02457511093881395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,128,0.004474666797452503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,16384,0.07229777839448717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,256,0.007091555330488417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,1024,0.004463110946946674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,8192,0.03453066613939073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,1536,0.037809779246648155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,12288,0.030890666776233252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,8192,0.037985778517193265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,2048,0.006724444528420766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,64,0.00816000004609426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,16384,0.05805244710710314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,256,0.00546577779783143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,5120,0.03607644306288825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,8192,0.0688284436861674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,3072,0.007585778004593319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,2048,0.008887999587588841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,128,0.002647111192345619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,8192,0.03234044379658169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,4096,0.021733333667119343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,3072,0.008867555194430882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,6144,0.03221244282192654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,3072,0.00790666706032223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,12288,0.09320177634557088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,128,0.004042666819360522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,128,0.0032311110860771606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,256,0.004118222329351637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,1536,0.01424533294306861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,3072,0.019303111566437613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,256,0.013381332986884646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,16384,0.09100355704625447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,2048,0.006508444332414203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,65536,0.06818399826685588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,1024,0.016985777351591323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,512,0.0032497776879204642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,7168,0.5501831372578939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,128,0.004320888883537716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,1024,0.005130666825506422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,7168,0.02722399930159251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,3072,0.016386666231685214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,12288,0.019134221805466544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,768,0.004927111168702443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,768,0.004156444635656145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,10240,0.0396071109506819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,8192,0.014315555493036905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,512,0.003458666718668408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,7168,0.010375111467308467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,64,0.005126222140259213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,6144,0.02383822202682495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,10240,0.04901955525080363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,6144,0.03271733389960395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,4096,0.2983946535322401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,32,0.0026951111439201566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,2560,0.022524444593323603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,12288,0.015114666687117683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,5120,0.01608711149957445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,5120,0.02234755622016059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,5120,0.008444444172912175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,3072,0.8783635563320584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,2560,0.027254222167862788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,5120,0.16949244340260824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,16384,0.08637066682179768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,4096,0.010961777634090848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,3072,0.0069955554273393415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,12288,0.01883555617597368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,1536,0.01038755559258991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,256,0.012873777912722694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,1024,0.00534400012758043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,7168,0.01183733344078064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,8192,0.014874666929244995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,768,0.00812177773978975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,768,0.0058577776783042485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,4096,0.7518106566535102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,2048,0.005221333354711533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,1536,0.03389599919319153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,1536,0.0069360000391801195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,2560,0.017415111263593037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,5120,0.056551112069023975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,2048,0.02676533328162299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,5120,0.008180444439252218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,1536,0.020024889045291476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,7168,0.011360000405046674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,65536,0.03919733232922024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,65536,0.13216622670491537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,4096,0.010619555910428366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,128,0.009639110830095079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,32,0.005038222091065513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,64,0.006007110906971826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,3072,0.1473004420598348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,2048,0.007863110966152614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,128,0.013610666824711693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,128,0.008192000289758047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,5120,0.009350222018029954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,128,0.012540444731712341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,7168,0.02826933397187127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,2048,0.012310222619109683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,512,0.0034968890249729156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,1536,0.18041955100165474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,2048,0.05330577823850843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,8192,0.01497244503762987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,3584,0.02012266715367635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,1536,0.01745244363943736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,1536,0.005179555465777715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,2048,0.005305777821275923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,128,0.002938666691382726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,12288,0.3852248986562093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,768,0.00796622203456031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,512,0.005418666534953647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,2048,0.01569422251648373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,5120,0.03090133269627889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,16384,0.026681777503755357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,65536,0.3642337852054172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,3072,0.007874666816658443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,768,0.006319111006127463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,3072,0.007576000359323289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,4096,0.016150222884284127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,2560,0.10129688845740424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,4096,0.009954666925801171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,128,0.007798222204049428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,4096,0.01590666671593984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,12288,0.030745777818891738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,3584,0.006806222101052602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,7168,0.023289778166347083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,8192,0.02362577782736884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,4096,0.027056889401541814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,2048,0.006133333262470033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,768,0.003863999827040566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,2048,0.008790222307046255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,12288,0.10405866967307197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,128,0.0041866666740841335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,3584,0.010237333675225576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,65536,3.7044410705566406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,8192,0.036329779360029436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,64,0.003481777591837777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,2048,0.013034666577974955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,512,0.013774221969975365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,64,0.00739555557568868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,10240,0.08382044235865276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,10240,0.01054133309258355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,768,0.0207315550910102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,2560,0.00665244460105896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,2048,0.0058613332609335584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,3072,0.01219288922018475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,2048,0.007051555646790399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,8192,0.033478223615222506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,7168,0.028239111105600994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,2560,0.017808000246683758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,65536,0.0761351121796502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,6144,0.009479111267460717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,1536,0.012643555800120035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,256,0.0052435555391841466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,4096,0.09207199679480658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,4096,0.040756444136301674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,10240,0.020822222034136455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,6144,0.008298666940795051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,2048,0.013794666363133324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,32,0.005510222166776657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,8192,0.01736088924937778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,16384,0.434641785091824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,768,0.0614373352792528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,16384,0.046576887369155884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,256,0.0032631111227803757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,2048,0.007490666376219855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,7168,0.009233777721722921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,8192,0.04203644394874573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,4096,0.0539902216858334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,768,0.02733777794573042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,8192,0.013711111413107978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,32,0.002862222285734283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,4096,0.006753777878151999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,512,0.02086044516828325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,7168,0.00945688866906696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,8192,0.012030222349696688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,3584,0.013269333375824822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,3072,0.017849778135617573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,4096,0.011284444067213269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,3584,0.010824888944625854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,12288,0.054142223464118115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,6144,0.021936888496081035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,12288,0.091648002465566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,32,0.0028373334142896864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,512,0.005088888936572605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,512,0.007376888559924231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,512,0.005392888767851724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,64,0.005173333403136996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,2048,0.008290666672918532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,2560,0.013005333642164866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,12288,0.04235466652446323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,2560,0.012311110893885294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,5120,0.008622222476535374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,6144,0.006491555521885554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,16384,0.04081688986884223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,1536,0.02698755595419142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,10240,0.02974489000108507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,3584,0.009317333499590555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,512,0.012662221988042196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,8192,0.04937866661283705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,32,0.006632888896597757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,1024,0.005114666703674528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,6144,0.011651555697123209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,7168,0.017465778523021273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,32,0.004951999833186467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,768,0.005399111244413588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,1536,0.009392889009581672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,7168,0.020576000213623047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,64,0.008216888540320927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,2048,0.025923555095990498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,64,0.0028613333900769553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,6144,0.020018666982650757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,768,0.005199999858935674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,3584,0.006839111033413146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,65536,0.2855795489417182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,32,0.005201777650250329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,1024,0.005097777893145879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,64,0.0099982221921285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,2048,0.007163555257850223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,7168,0.046260442998674184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,8192,0.04007822275161743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,4096,0.12206844488779704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,512,0.023780445257822674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,3584,0.013656888571050433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,16384,0.13150400585598415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,10240,0.12388000223371719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,1024,0.0064897777305708984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,256,0.007137777904669444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,6144,0.1935537788603041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,128,0.03350933392842611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,3584,0.017937777770890128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,2048,0.008076444268226624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,16384,0.10194400284025405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,7168,0.10428355799780951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,1536,0.011261333194043903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,12288,0.028592000404993694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,512,0.004297777596447202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,8192,0.012551111479600271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,256,0.013818666338920593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,768,0.003950222084919612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,5120,0.02759733299414317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,5120,0.02088444431622823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,256,0.005479111025730769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,4096,0.012545777691735161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,6144,0.01090222183201048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,3072,0.0239964442120658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,10240,0.008600888980759515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,1536,0.006643555644485686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,512,0.015621332658661736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,1536,0.006625777731339137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,128,0.005373333477311664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,128,0.0037208890749348533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,12288,0.04046666622161865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,7168,0.33761244350009495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,65536,0.026339555780092876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,1024,0.022240888741281297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,5120,0.013342222405804528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,64,0.03488711184925503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,65536,0.19795644283294678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,2560,0.014997333288192749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,64,0.002757333426011933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,512,0.025011556016074285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,768,0.005423111220200856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,6144,0.03958844476275974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,1536,0.010063110954231685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,65536,0.5517413351270888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,512,0.049154665735032826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,256,0.46039819717407227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,4096,0.006633777585294511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,6144,0.011624000138706632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,32,0.0052977779673205475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,64,0.0028515555378463534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,512,0.0032968889507982465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,4096,0.07192622290717231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,3584,0.03022222386466132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,2560,0.010232888989978367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,4096,0.029359999630186293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,16384,0.7845599916246203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,768,0.004496888981925117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,2048,0.01992266707950168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,4096,0.11727733082241482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,10240,0.00963644435008367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,32,0.0026346666531430352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,256,0.008553778131802877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,3072,0.008063111040327284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,8192,0.36568533049689395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,6144,0.00904888908068339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,16384,0.1649751133388943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,768,0.0047484445903036325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,10240,0.013291555146376291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,2048,0.008292444050312042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,16384,0.5164995723300511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,10240,0.061853335963355176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,2048,0.007844444778230455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,10240,0.015899555550681222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,3584,0.03174133433236016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,2560,0.007143999967310164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,768,0.003810666501522064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,2560,0.01385600037044949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,64,0.0026400000270869997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,1024,0.006071110980378256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,3584,0.007560888926188151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,1024,0.007605333295133378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,64,0.004753777964247597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,6144,0.0058248887459437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,7168,0.2675422297583686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,768,0.005471999860472149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,10240,0.018220444520314533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,12288,0.0928693347507053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,32,0.11610489421420628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,32,0.0027822222974565295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,3072,0.025065778030289546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,128,0.004512888689835866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,2048,0.3160444365607368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,512,0.03163555595609877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,256,0.004937777916590373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,4096,0.017483555608325534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,3072,0.010175111393133799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,4096,0.02836622132195367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,64,0.01074399964676963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,12288,0.006976889239417181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,768,0.0047288888858424295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,2048,0.01865866614712609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,16384,0.02257599929968516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,1536,0.050513777467939586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,8192,0.10863644546932644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,65536,0.09240977631674872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,768,0.07311644156773885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,7168,0.009731555150614845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,4096,0.009408888717492422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,65536,1.7753119998508031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,2560,0.00944088896115621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,6144,0.027021333575248718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,2048,0.008472888833946651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,12288,0.045291556252373584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,6144,0.0606879989306132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,512,0.007648000286685095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,4096,0.014864888456132678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,128,0.0053431110249625305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,2560,0.009083555804358589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,512,0.003660444584157732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,1024,0.013648889131016202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,128,0.011010666688283285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,1536,0.03779022230042352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,4096,0.018889778190188937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,3584,0.08283199866612752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,8192,0.00889244427283605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,256,0.004583999928500918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,3072,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,64,0.0027217778066794076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,3584,0.029300444655948218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,4096,0.0138568886452251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,16384,0.04349422123697069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,1536,0.009744000103738572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,6144,0.14793866210513645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,256,0.0033164444482988785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,16384,0.03380533390574985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,7168,0.15300977230072021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,6144,0.07908355527453952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,4096,0.009876444107956356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,3584,0.058232890235053174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,6144,0.011016888750924004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,3584,0.010423999693658618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,8192,0.01204355557759603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,5120,0.008775110873911116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,512,0.0047484445903036325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,8192,0.028675556182861328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,32,0.005599110904667113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,768,0.005514666852023866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,8192,0.01684799955950843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,5120,0.023781332704755995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,64,0.00600622221827507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,2048,0.013379555609491138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,128,0.003822222352027893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,64,0.0028159999185138275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,512,0.005600000007285013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,5120,0.010078222387366826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,2048,0.04017511010169983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,7168,0.039312889178593956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,65536,0.06956622335645887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,6144,0.006219555520349079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,768,0.19002577993604872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,768,0.004688888788223267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,2048,0.012663111090660095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,6144,0.043122665749655835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,3584,5.929481082492405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,128,0.004289777742491828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,1536,0.011494222614500256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,65536,0.193740447362264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,6144,0.046262221203909985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,8192,0.015051555302408008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,768,0.006774222271309958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,2560,0.009221333596441481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,3072,0.05355199840333727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,5120,0.018911111685964797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,768,0.008244444098737504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,6144,0.01923022170861562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,16384,0.6415680249532064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,8192,0.02702399922741784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,64,0.005708444449636672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,2048,0.1634311146206326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,3072,0.009942221972677443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,5120,0.028134223487642076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,6144,0.007094221810499827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,3072,0.009848000274764167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,10240,0.03254755669169956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,1024,0.007228444847795699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,10240,0.012988444831636218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,5120,0.15730488300323486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,8192,0.18063110775417754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,2560,0.007851555943489075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,256,0.0037546666959921518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,64,0.04013333386845059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,512,0.0075706665714581805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,128,0.008664888640244802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,16384,0.03296711047490438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,4096,0.00813155538505978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,1024,0.0058319999112023254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,128,0.026532444689008925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,512,0.011887110769748688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,1024,0.006531555619504716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,10240,0.03491733471552531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,8192,0.013559111290507846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,1024,0.010263111028406356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,768,0.04533511069085863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,4096,0.008567999634477828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,10240,0.0177528891298506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,768,0.007681777907742395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,4096,0.005613333235184352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,768,0.022112000319692824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,16384,0.1034968892733256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,768,0.004791111167934206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,3072,0.02921688887808058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,3072,0.013032000097963544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,256,0.00555377784702513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,6144,0.5916764471266005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,3072,0.27375910017225474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,768,0.03980977667702569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,3584,0.025546666648652818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,128,0.003168888803985384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,16384,0.43242488967047793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,16384,0.01937244501378801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,16384,0.02903377678659227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,12288,0.021751999855041504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,3072,0.078146669599745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,5120,0.07230666610929701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,256,0.005643555687533484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,3584,0.011176889141400656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,5120,1.277801725599501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,512,0.008303999900817871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,65536,0.14044000042809382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,1536,0.007048889166778988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,2560,0.010125333236323463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,8192,0.009482666850090027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,4096,0.0259226659933726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,2560,0.018648000227080453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,10240,0.006807111203670502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,256,0.007520000139872233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,128,0.005173333403136996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,12288,0.019156444403860304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,32,0.002570666579736604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,10240,0.01036088913679123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,16384,0.5010879834493002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,2048,0.014540443817774454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,128,0.004132444245947732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,4096,0.008016000191370646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,64,0.005404444618357553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,5120,0.06955733564164904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,32,0.0033333332588275275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,128,0.008975110948085785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,16384,0.02823822365866767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,1536,0.029952890343136255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,10240,0.022484444909625586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,2048,0.011799111134476133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,3072,0.015073777900801765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,3584,0.008278222547637092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,1024,0.006852444261312485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,2560,0.0122079998254776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,65536,0.3251333236694336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,64,0.0063893331421746155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,10240,0.12983911567264134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,64,0.0028799999919202593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,4096,0.010144000252087912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,5120,0.026103110777007207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,128,0.007060444189442529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,3072,0.015019555886586508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,1536,0.007802666889296637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,10240,0.017503110898865592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,768,0.004974222017659081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,4096,0.006187555690606435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,128,0.003192000091075897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,1536,0.008089777496125963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,3584,0.018852444158660043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,768,0.021773333350817364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,10240,0.13229421774546304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,1536,0.011114666859308878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,3584,0.15691288312276205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,8192,0.011774222056070963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,1024,0.011233777635627322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,7168,0.09717244572109646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,768,0.003927999900446998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,16384,0.0210844443904029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,1024,0.0038444445365005066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,2560,0.01425333321094513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,2560,0.060476442178090416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,2560,0.00871733327706655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,3584,0.00795199970404307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,512,0.019931554794311523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,16384,0.08667200141482884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,1536,0.005026666654480828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,16384,0.053230222728517324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,16384,0.12020622359381782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,5120,0.022202667262819078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,3072,0.006822222222884496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,7168,0.013067555924256643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,2560,0.009840000006887648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,256,0.00648266656531228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,3072,0.3875324461195204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,7168,0.011564444336626263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,6144,0.10575466685824925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,1536,0.004953777624501122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,64,0.12960533301035562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,64,0.011671110987663269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,65536,0.2842000060611301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,12288,0.0169813334941864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,4096,0.02611911131276025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,8192,0.05244266655710009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,2560,0.013743999931547375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,8192,0.030751999881532457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,8192,0.24794310993618437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,1536,0.036008887820773654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,64,0.00554666668176651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,12288,0.04690044456058078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,65536,0.16643467214372423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,128,0.039826667971081205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,2048,0.029198222690158423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,2048,0.1109608875380622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,10240,0.11248532931009929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,2560,0.05299199952019585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,12288,0.0173999998304579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,1024,0.0402817792362637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,768,0.005956444475385878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,65536,0.6235235532124838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,2048,0.005942222144868638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,32,0.002941333378354708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,2048,0.017367111312018502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,7168,0.014558222558763293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,8192,0.021408889028761122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,8192,0.038423111041386925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,512,0.007080000307824876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,12288,0.023668444818920557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,256,0.03834311167399088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,4096,0.037745777103636004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,65536,0.17230667008293998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,2048,0.008014221986134848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,64,0.005220444252093633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,7168,0.014696889453464083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,8192,0.011779555843936073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,3584,0.006681777950790193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,32,0.02414400047726101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,512,0.007284444239404466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,6144,0.011457777685589261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,2560,0.011326221956147088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,1536,0.004761777818202972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,64,0.0034764446318149567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,64,0.006362666686375936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,3072,0.02920177910063002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,2560,0.01053955571519004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,768,0.0036951113078329298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,256,0.004392000122202767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,8192,0.032984889215893216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,32,0.005150222116046482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,8192,0.7773653136359321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,6144,0.03730488816897074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,2048,0.009806222385830348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,16384,0.042295998997158475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,16384,0.06812444660398695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,3584,0.037984887758890785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,5120,0.008835555778609382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,16384,0.020584000481499564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,512,0.003504000190231535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,7168,0.0336480008231269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,1024,0.0052044445441828836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,10240,0.726436456044515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,64,0.007263110743628607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,768,0.025176889366573755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,10240,0.30422843827141655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,256,0.013424889081054263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,3584,0.036473777559068464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,16384,0.024911999702453613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,3584,0.11170132954915364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,4096,0.06239910920461019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,3072,0.02499466637770335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,512,0.0064195555945237475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,12288,0.1302328904469808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,1024,0.8195404476589627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,5120,0.01514755520555708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,3072,0.0688284436861674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,768,0.007148444652557373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,2048,0.008362666600280339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,5120,0.015289778510729471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,1536,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,16384,0.025376889440748427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,128,0.005850666513045629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,5120,0.015034667319721647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,1024,0.022738666998015508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,512,0.005176888985766305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,8192,0.011561777856614856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,12288,0.032955557107925415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,1024,0.013181333740552267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,10240,0.006437333507670297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,3584,0.05236000153753492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,16384,0.16412711143493652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,4096,0.005852444304360284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,5120,0.015781333049138386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,65536,1.4139742321438258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,12288,0.040439999765819974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,65536,0.014948444234000312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,16384,0.03848177856869168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,65536,0.010504888991514841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,8192,0.011400000088744693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,1024,0.023524444964196947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,6144,0.03568799959288703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,12288,0.041823999749289624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,256,0.04199466771549649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,256,0.0047893333766195504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,768,0.029940442906485662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,8192,0.014592889282438489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,5120,0.020672889219390023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,1536,0.2979164388444689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,3072,0.011387555963463254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,1024,0.01258933295806249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,8192,0.04583911101023356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,768,0.005945777727497949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,2048,0.009232000344329411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,4096,0.006193777753247156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,2560,0.05822133355670505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,2048,0.0209840006298489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,12288,0.01629333363638984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,2048,0.01348533315791024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,512,0.005466666486528184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,1024,0.014948444234000312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,32,0.005240889059172736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,128,0.0069253332912921906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,2560,0.23248444663153756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,7168,0.020785777105225455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,2560,0.01925511161486308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,128,0.005101333475775189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,16384,0.02184711065557268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,3584,0.010952889091438718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,6144,0.0203475554784139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,2560,0.013265777793195514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,32,0.0026844443960322272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,12288,0.025270221961869135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,4096,0.012230222423871359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,12288,0.02238399949338701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,256,0.003140444349911478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,65536,0.5359431372748481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,768,0.006748444504208035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,8192,0.015407111909654407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,4096,0.02887644370396932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,65536,0.07492622401979235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,65536,1.1787182490030925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,1536,0.007084444165229797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,10240,0.022867555419603985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,2560,0.015507555670208402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,5120,0.011112000379297467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,128,0.004084444294373195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,12288,0.058133330610063344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,2560,0.007703999678293864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,256,0.006805333412355847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,7168,0.0397093329164717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,32,0.002781333401799202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,2560,0.011393778026103973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,8192,0.025877333349651758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,1024,0.008933333059151968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,1024,0.005048000150256687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,3072,0.0069466663731469055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,65536,0.18435466289520264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,1536,0.012073777616024017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,12288,0.19632266627417672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,128,0.006072000082996156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,3072,0.026520000563727483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,12288,4.944426642523871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,10240,0.1844115522172716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,7168,0.014279110564125909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,1024,0.005717333406209946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,1024,0.027486221657858953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,16384,0.019620445039537217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,6144,0.009890666438473595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,2048,0.07510844204160902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,16384,0.050642665889528066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,16384,0.058373332023620605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,2560,0.06879111131032307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,768,0.006779555645253923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,128,0.00508977762526936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,8192,0.011682666838169098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,256,0.006545777950021956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,1024,0.005127999931573868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,65536,7.3896730211046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,1536,0.008550222549173567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,3584,0.01982311076588101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,12288,0.018170666363504197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,32,0.002959111084540685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,12288,0.1715395583046807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,7168,0.029436445898479883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,256,0.03602133194605509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,512,0.08357777860429551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,7168,0.027645332945717707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,10240,0.13394578297932944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,768,0.005595555735958948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,12288,0.04434577624003092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,10240,0.027922666735119287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,6144,0.018258665998776753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,256,0.008979555633332994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,256,0.003160888950030009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,128,0.0038159998754660287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,16384,0.05268088976542155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,128,0.002842666581273079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,128,0.007989333735571967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,4096,0.03640800052218967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,4096,0.03049066662788391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,64,0.007333333293596904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,7168,0.011758222348160215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,1024,0.005324444423119227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,3584,0.007841777470376756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,768,0.014967999524540372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,12288,0.054339554574754506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,3072,0.007511110769377814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,256,0.00850399997499254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,12288,0.07778577672110663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,10240,0.8244178030225965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,1024,0.020626667473051283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,16384,0.05703022082646688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,2048,0.011776888536082374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,256,0.006713777780532837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,64,0.00703288863102595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,6144,0.006153777655627992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,768,0.01351999988158544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,3584,0.016536888149049547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,256,0.005474666754404704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,8192,0.008897778060701158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,65536,0.11604444185892741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,128,0.007075555622577667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,3072,0.015581332974963717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,6144,0.016950221525298226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,6144,0.013368888861603208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,4096,0.007896888587209914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,10240,0.05577955643335978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,12288,0.15732622146606445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,64,0.002732444347606765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,3584,0.011919111013412476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,32,0.005960888746711943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,512,0.014903111590279473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,1024,0.00519822206762102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,128,0.004254222330119875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,768,0.017158221867349412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,10240,0.02737866673204634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,3072,0.008815999660227034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,3072,0.019765333996878732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,8192,0.021274665991465252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,4096,0.008365333080291748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,1536,0.024718221690919664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,16384,0.024528000089857314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,4096,0.036706666151682533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,7168,1.7156640158759222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,1536,0.007778666913509369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,65536,1.1836782031589084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,3072,0.013743999931547375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,10240,0.27876443333095974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,512,0.005023111071851519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,3584,0.09041066964467366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,10240,0.21682843897077772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,4096,0.013329777452680798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,6144,0.01755822201569875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,3072,0.03885511226124234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,7168,0.03549244337611728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,16384,3.3057867685953775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,3584,0.018434666925006442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,16384,0.024672889047198828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,8192,0.033747557136747576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,6144,0.07420888874265882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,1024,0.007061333292060428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,1024,0.007094221810499827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,4096,0.022557333111763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,5120,0.028377778000301782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,1536,0.05008888906902737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,128,0.00371111101574368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,768,0.011346666349305047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,256,0.026920888159010146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,7168,0.18470488654242623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,2560,0.009378666679064432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,1536,0.030864887767367895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,512,0.005079111291302575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,8192,0.006544000158707301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,5120,0.020640000700950623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,1536,0.1283244424396091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,4096,0.009469333622190688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,1536,0.005987555616431766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,2048,0.01569422251648373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,512,0.005114666703674528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,65536,0.03490133417977227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,512,0.0041857775714662345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,2048,0.006105777704053455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,256,0.021578666236665513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,7168,0.046554668082131274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,64,0.0753928886519538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,2560,0.17782310644785562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,1536,0.008834666675991481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,5120,0.5128400060865614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,7168,0.6663537555270725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,2048,0.0073902226156658595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,12288,0.1884231170018514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,8192,0.03120622369978163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,2560,0.0068764446510208984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,12288,0.12147288852267796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,2560,0.015303111738628812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,2560,0.027264000640975103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,1536,0.010812444819344414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,256,0.004228444563017951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,2048,0.011720889144473605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,8192,0.02693422304259406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,12288,0.3979493247138129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,4096,0.006511999915043513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,6144,1.2937208811442058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,10240,0.037045333120558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,1024,0.00959822204377916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,3584,0.010296000374688042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,6144,0.08717244201236302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,6144,0.20056711302863228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,256,0.004370666626426908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,5120,0.007087111473083496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,1024,0.004348444441954295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,3072,0.010080888867378235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,6144,0.00997511131895913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,3072,0.006726222319735422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,1536,0.0096124443742964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,128,0.006248888870080312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,12288,0.09587111075719197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,256,0.022992889086405437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,8192,0.046712001164754234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,16384,0.07173511054780748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,2048,0.010213333699438307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,1024,0.05068711108631558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,3584,0.01274311128589842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,8192,0.012348444097571902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,6144,0.060812446806165904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,32,0.0033004445334275565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,2560,0.01789777808719211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,128,0.006009777800904379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,128,0.003318222239613533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,768,0.004256000121434529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,7168,0.017782222893502977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,3072,0.018416888184017606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,3584,0.008825778133339351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,2048,0.025358221597141687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,256,0.006079111248254776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,256,0.005327111317051782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,16384,2.187500423855252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,2048,0.06580888562732272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,2560,0.010761777559916178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,12288,0.026230222649044458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,2048,0.00849777791235182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,16384,0.268958224190606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,2560,2.1057590908474393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,10240,0.014922666880819531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,32,0.021697777840826247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,1024,0.07614755630493164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,128,0.04974933465321859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,2560,0.007343110938866933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,8192,0.2047599951426188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,1024,0.04827733172310723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,3584,0.02635555631584591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,3072,0.006831999868154526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,12288,0.009184000392754873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,3072,0.023471999499532912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,65536,0.06749066379335192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,32,0.0028977776981062363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,7168,0.016695999436908297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,768,0.005305777821275923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,4096,0.07862044705284967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,3584,0.018294221825069852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,12288,0.1499662266837226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,128,0.049217777119742505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,5120,0.02460711035463545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,32,0.01075111081202825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,7168,0.026947556270493403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,4096,0.012655999925401477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,10240,0.015830222103330824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,3584,0.007072000039948358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,128,0.008450667063395182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,5120,0.0450853341155582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,768,0.008825778133339351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,2560,0.013854222165213691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,1536,0.007800000409285228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,5120,0.08951289123959011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,768,0.008959111240175035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,16384,0.058740443653530545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,64,0.008455110920800103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,2048,0.030795554320017498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,10240,0.4468666712443034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,7168,0.030051555898454454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,1024,0.004111111164093018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,65536,0.20715644624498156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,3072,0.00683111117945777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,65536,0.22994311650594076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,3584,0.006934222247865465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,12288,0.015576000014940897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,2048,0.018912000788582694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,12288,0.1106497777832879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,768,0.005921777751710679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,65536,0.14984267287784153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,8192,0.10380888647503322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,8192,0.9192231496175131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,65536,2.2765013376871743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,16384,0.09561244646708171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,64,0.012906666431162091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,16384,0.06508088774151273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,12288,0.054323554039001465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,2560,0.017116443978415597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,16384,0.05934311283959282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,12288,0.03384000062942505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,1536,0.012259555359681448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,2560,0.00941777808798684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,512,0.004551110996140374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,3584,0.006743999818960826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,6144,0.01350488927629259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,1536,0.010122666756312052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,2048,0.006862222320503659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,65536,0.23745422893100313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,12288,0.042747557163238525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,256,0.0069306666652361555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,2048,0.012996444271670448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,3072,0.006879999819729064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,3072,0.007232888705200619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,3072,0.006912888752089606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,12288,0.025640888346566096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,8192,0.017441777719391715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,2048,0.006604444649484422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,8192,0.011568889021873474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,3584,0.011696000066068439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,3584,0.00886311133702596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,65536,0.6202889018588597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,4096,0.020304888486862183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,10240,0.010891555911964841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,3584,0.009528888596428765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,10240,0.07537244425879584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,768,0.003564444267087512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,7168,0.016057777735922072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,10240,0.022610666023360357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,128,0.004401777767472797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,768,0.0409804450141059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,12288,0.025469332933425903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,768,0.008904000123341879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,12288,0.08008888694975111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,5120,0.22367734379238555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,3584,0.014102222190962898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,1024,0.008240888516108194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,4096,0.01463644372092353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,2048,0.010958222051461538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,768,0.0041244443919923575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,1024,0.020785777105225455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,512,0.00702133360836241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,3584,0.010149333212110732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,65536,0.07937244574228923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,6144,0.2976337803734673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,10240,0.006958222223652734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,2048,0.03784088955985175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,32,0.007282666862010956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,3584,0.11160622702704535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,65536,0.11202400260501438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,7168,0.006507555643717448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,6144,0.009708444277445475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,4096,0.07788978020350139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,256,0.00517866677708096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,4096,0.007814221911960179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,3584,0.011771555576059552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,64,0.006745777610275481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,2560,0.009806222385830348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,10240,0.031089779403474595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,768,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,10240,0.0529022216796875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,256,0.0070879997478591064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,128,0.006887110984987683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,16384,0.020152888364262052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,6144,0.059608002503712974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,768,0.007048000064161088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,10240,0.5310266812642416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,12288,0.017639999588330586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,10240,0.023932443724738225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,2048,0.014557333456145393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,768,0.004967111266321606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,1536,0.020505777663654752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,64,0.003551111039188173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,256,0.0034808889031410217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,2048,0.07460444503360324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,16384,0.16979732778337267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,3584,0.014800000521871777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,2048,0.009272889130645329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,1536,0.006745777610275481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,768,0.005896889087226655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,1536,0.017314665847354464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,8192,0.012355555262830524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,128,0.006182222316662471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,12288,0.010199111368921068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,8192,0.1123431126276652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,64,0.006011555592219035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,3072,0.0058648888435628675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,32,0.0037066667444176147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,2560,0.023737778266270954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,128,0.008103110724025303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,65536,0.2386782169342041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,2048,0.00850577818022834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,65536,0.1836328903834025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,128,0.005000888887378904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,65536,0.6879546907213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,16384,0.24731734063890246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,5120,0.019488889310095046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,8192,0.021365332934591506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,64,0.005434666656785541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,8192,0.01568000018596649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,6144,0.024626667300860088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,32,0.002887111157178879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,3584,0.07193422317504883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,1024,0.0074879998962084455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,7168,0.020399999287393358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,64,0.01162577751610014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,2048,0.01680888897842831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,64,0.002642666714058982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,2048,0.025440000825458105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,5120,0.011757333245542316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,8192,0.03064089020093282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,8192,0.006478222293986215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,1536,0.006320000108745363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,128,0.009956444303194681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,3072,0.016386666231685214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,1024,0.0052586665583981415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,3072,0.02312266661061181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,65536,0.09676800171534221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,256,0.01035644445154402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,4096,0.04728088776270548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,3584,0.14816266960567898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,512,0.004685333205593957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,2560,0.22868445184495714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,512,0.005389333185222413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,2560,0.05085866649945577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,3584,0.14517154958513048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,5120,0.014719999498791166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,4096,0.014768888552983602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,2560,0.07133244143591987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,3072,0.009992000129487779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,1536,0.06861866845024957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,3584,0.07393866777420044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,2560,0.007227555745177799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,5120,0.04750488864050972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,2560,0.007092444433106317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,2560,0.007822222179836696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,128,0.0053022222386466126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,1536,0.07438044415579902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,3072,0.01074488874938753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,768,0.0058328890138202244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,3072,0.03413866625891791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,32,0.0028986665937635633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,128,0.01183377785815133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,512,0.007532444265153673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,256,0.009823111196359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,16384,0.030607110924190942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,7168,0.4805475340949164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,6144,0.02864266766442193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,10240,0.01704088846842448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,1024,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,3072,0.010091555615266165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,16384,0.1546880006790161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,1536,0.1409697797563341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,256,0.007854222423500484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,1024,0.004964444372389051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,1024,0.00647199981742435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,1024,0.025655110677083332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,768,0.07727111048168607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,2560,0.34597688251071507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,768,0.0040995557275083335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,5120,0.013560000393125745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,768,0.014678221609857349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,3584,0.009494221872753566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,1024,0.006200888918505774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,65536,8.922720167371962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,128,0.019301333361201815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,2048,0.011307555768224927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,32,0.00904177791542477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,1024,0.011754666765530905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,6144,0.07409244113498263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,6144,0.0233288887474272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,6144,0.05612533622317844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,128,0.005398222141795688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,8192,0.044665777020984225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,512,0.046875556310017906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,10240,0.0063902222447925145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,12288,0.23144533899095324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,6144,0.014491554763582019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,8192,0.3003466659122043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,10240,0.01679733395576477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,3072,0.03180888957447476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,65536,0.08473777770996094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,4096,0.010891555911964841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,10240,0.12888177235921225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,512,0.005066666752099991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,1536,0.005191111316283544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,512,0.007416889071464539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,32,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,1536,0.006550222221348021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,16384,0.05528444382879469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,5120,0.006478222293986215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,512,0.00656444455186526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,1024,0.07033955388598971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,32,0.004671999977694618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,16384,0.10494399733013576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,10240,0.05691288577185737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,1536,0.005607111172543631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,512,0.0075093333919843035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,2560,0.019176888797018263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,12288,0.010788444843557147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,128,0.0034933334423436057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,12288,0.016511999898486666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,5120,0.026196445027987163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,1024,0.0057964444988303715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,65536,0.07570844226413302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,7168,0.02791555391417609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,5120,0.02488888965712653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,64,0.004975999808973736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,128,0.004703999807437261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,32,0.006736889067623351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,64,0.005445333404673471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,12288,0.22057778305477568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,4096,0.00978488889005449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,512,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,12288,0.014952888091405233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,3584,0.007418666448858049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,64,0.0026684444811609057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,512,0.08677866723802354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,10240,0.11104533407423232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,8192,0.01944177846113841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,256,0.003499555504984326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,65536,0.48721954557630753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,2560,0.013232000172138214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,8192,0.1419582234488593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,8192,0.07870844337675306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,2048,0.012111999922328524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,4096,0.006409777949253718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,8192,0.029213332467608984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,3584,0.013587555951542325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,2048,0.00721866637468338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,768,0.007792000141408708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,64,0.006495999793211619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,4096,0.009197333620654212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,512,0.0068488890926043195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,16384,0.7249573071797689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,32,0.003669333126809862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,128,0.003321777822242843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,2048,0.018838221828142803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,32,0.003024000053604444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,128,0.014172444740931192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,128,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,768,0.006895111252864202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,65536,0.3090133402082655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,5120,0.025252444876564875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,1536,0.00675644435816341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,4096,0.317575110329522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,1024,0.007302222152551015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,1024,0.00620977787507905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,6144,0.033907555871539645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,1024,0.005790222022268508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,2048,0.007786666353543599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,2048,0.20112178060743544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,1024,0.006025777922736273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,10240,0.06541955471038818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,512,0.059587558110555015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,64,0.010760000182522668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,32,0.005255110975768831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,1024,0.006969778074158563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,4096,0.03819466630617777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,65536,0.40102399720085996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,1024,0.004074666649103165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,7168,0.02348622183005015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,7168,3.0365447998046875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,10240,0.019555555449591745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,2048,0.01166133334239324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,768,0.03750844465361701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,3584,0.00885688927438524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,256,0.0047217777205838096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,4096,0.29049955474005806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,12288,0.026177777184380427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,1024,0.006015111174848344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,5120,0.009293333523803288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,32,0.006404444575309753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,5120,0.2441928916507297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,256,0.00342755557762252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,16384,0.018977777825461496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,1024,0.004121777911980947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,1536,0.02215555475817786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,2048,0.005345777918895085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,10240,0.06471911403867933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,10240,2.0958436330159507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,128,0.0034346667428811393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,5120,0.00790666706032223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,3072,0.010093332992659675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,3584,0.009145778086450364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,768,0.007612444460391998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,7168,0.028632889191309612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,768,0.0053013331360287136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,2560,0.015584000282817416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,4096,0.008181333541870117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,8192,0.011483555866612328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,65536,0.06811644633611043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,1024,0.016880000631014507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,64,0.007381333245171442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,8192,0.010078222387366826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,1024,0.005583999885453119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,5120,0.010828444527255164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,8192,0.05903999672995674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,4096,0.027821332216262817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,12288,0.026657778355810378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,16384,0.02179644505182902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,10240,0.12046044402652317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,1536,0.009094222552246517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,768,0.006532444308201472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,512,0.011191999746693505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,512,0.007261333366235097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,4096,0.011583111352390714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,32,0.008681777450773451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,4096,0.00849777791235182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,3584,0.021018667353524104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,6144,0.023971555961502924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,12288,0.015240000353919135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,8192,0.029179556502236262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,10240,0.016333333320087857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,128,0.009116444322797988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,2048,0.007177777588367462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,768,0.005616000129116907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,1024,0.006230222268237009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,6144,0.013191111385822296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,6144,0.021747555997636583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,16384,0.10188800096511841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,256,0.016524444023768108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,2048,0.014279999666743808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,7168,0.01767200065983666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,1024,0.005495999836259418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,1536,0.01611022154490153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,2560,0.029799110359615747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,2048,0.00665866666369968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,12288,0.6695626576741537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,7168,0.014192000031471252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,128,0.006058666855096817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,32,0.013045333325862885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,768,0.007738666401969061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,6144,0.01827022267712487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,2048,0.005902222047249476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,3584,0.009387555221716562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,256,0.006958222223652734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,32,0.00600888869828648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,64,0.0026275554878844153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,3072,0.15172799428304037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,16384,0.01993333299954732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,16384,0.023440000083711412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,8192,0.036512000693215266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,12288,0.01814311080508762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,256,0.0028435554769304064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,65536,0.3529813289642334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,7168,0.33506933848063153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,6144,0.07410489188300239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,5120,0.0634817745950487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,1024,0.040491554472181536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,256,0.01037333326207267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,2048,0.010515555739402771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,2560,0.010045333041085137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,12288,0.0096124443742964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,16384,0.05883110894097222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,16384,0.05368444323539734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,768,0.005743999862008625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,6144,0.04943022131919861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,8192,0.024719999896155462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,6144,0.0063902222447925145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,4096,0.011321778098742167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,6144,0.009979555176364051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,64,0.0026613333159022858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,1024,0.005249777601824866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,6144,0.3383315669165717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,10240,0.038549333810806274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,6144,0.023919110496838886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,5120,0.010073777702119615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,512,0.004471111214823193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,7168,0.013539555172125498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,128,0.004998222407367495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,2560,0.0070159998204973005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,8192,0.023066666391160753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,512,0.007140444384680853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,65536,0.052412443690829806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,7168,0.029166221618652344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,2048,0.04087288843260871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,3584,0.01715555621518029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,65536,0.13940710491604275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,32,0.012116444607575735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,5120,0.011973333027627734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,1024,0.006271111054552927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,256,0.011725333001878528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,65536,0.060773332913716636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,2560,0.012240889171759287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,2560,0.008815111385451423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,2048,0.00665155549844106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,10240,0.11185777187347412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,1536,0.008231110870838165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,2560,0.01575288838810391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,5120,0.8396222326490613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,32,0.0069679998689227635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,1536,0.37958489523993594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,5120,0.033328887489106916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,512,0.004293333325121137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,8192,0.028872887293497723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,65536,0.07431555456585355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,3584,0.021015110943052504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,5120,0.009107555780145857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,3584,0.025813332862324182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,2560,0.006864000111818314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,1024,0.008835555778609382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,768,0.01516533394654592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,8192,0.009858667022652095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,3072,0.05989333656099108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,4096,0.00775111135509279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,10240,0.03920533259709676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,10240,0.01792622274822659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,1536,0.009176000124878353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,1024,0.007686222592989604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,128,0.007268444531493717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,64,0.00554933316177792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,5120,0.028314666615592107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,1024,0.003949333396222857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,65536,0.2877635425991482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,6144,0.01723911033736335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,512,0.007283555964628856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,768,0.012030222349696688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,12288,1.1336346732245552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,2560,0.1072853340042962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,4096,0.01181422256761127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,3072,0.03273333443535699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,65536,0.3785795635647244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,7168,0.00868800034125646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,512,0.011007111105653973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,10240,0.06731733348634508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,16384,0.4409964349534776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,1024,0.010302222437328762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,3072,0.10551022158728705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,12288,0.04149511125352647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,8192,0.006100444330109491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,8192,0.012846222354306115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,5120,0.016632889707883198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,6144,0.05499644411934746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,8192,0.04319644305441114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,6144,0.018740443719757926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,5120,0.021000888612535264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,7168,0.029694222741656836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,2048,0.013241777817408243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,7168,0.022992889086405437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,512,0.0052488889131281115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,256,0.0057608890864584185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,1536,0.019816888703240287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,8192,0.17379911740620932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,512,0.004599110947714912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,1024,0.005992888990375731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,6144,0.010667555862002902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,7168,0.05545777744717068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,7168,0.013487111363146039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,1024,0.0041955556306574075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,256,0.0047111109726958806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,8192,0.013989332649442883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,1024,0.0158942225906584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,256,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,4096,0.07308710945977105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,5120,0.009180444810125563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,16384,0.009694221946928237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,512,0.005582222094138463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,4096,0.01146488885084788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,512,0.01644000079896715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,1024,0.006483555667930179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,1024,0.003920888735188378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,5120,0.024029332730505202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,7168,0.0064444442590077715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,16384,0.034224887688954674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,512,0.0032755554550223877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,6144,0.009516444471147325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,128,0.004052444464630551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,16384,0.017324444320466783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,128,0.002847111059559716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,65536,0.2156053384145101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,2048,0.005400889035728242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,32,0.0024888888001441956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,2560,0.007532444265153673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,4096,0.007689778175618913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,768,0.05281244383917915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,1024,0.03932266765170627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,16384,0.08555733495288426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,16384,0.14325777689615884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,12288,0.035275555319256253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,65536,0.016700444949997794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,8192,0.014197332991494073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,512,0.0058133333093590206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,512,0.006103111224042044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,4096,0.014565333724021912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,12288,0.02012800011369917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,65536,0.10009155670801799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,8192,0.006728000111050076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,64,0.00889955543809467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,6144,0.1342595550749037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,6144,0.13358578417036268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,2048,0.006281777802440856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,7168,0.008536888493431939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,768,0.01670222149954902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,65536,0.10605244504080878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,8192,0.1412737766901652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,4096,0.019281778070661757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,2048,0.007252444823582967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,64,0.019318222999572754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,10240,0.337391111585829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,12288,0.08311466375986735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,3072,0.011431111229790581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,6144,0.01445777714252472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,7168,0.019703111714786954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,2048,0.018917333748605516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,5120,0.022446221775478784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,64,0.020900444851981267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,512,0.003607999947335985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,4096,0.008273777862389883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,256,0.0027991111079851785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,5120,0.023874666955735948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,7168,0.014273777604103088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,16384,0.10216622220145331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,10240,0.11482311619652642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,5120,0.018624888526068795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,65536,0.48613866170247394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,65536,0.023960888385772705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,128,0.006272888845867581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,4096,0.005917333480384614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,6144,0.18336978223588732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,8192,0.008721777962313758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,64,0.00600355573826366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,7168,0.02293955617480808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,16384,0.021337777376174927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,6144,0.02387111054526435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,2560,0.006123555617200003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,32,0.004929777648713854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,10240,0.014999111493428549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,64,0.024953777591387432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,12288,0.018096889058748882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,768,0.005324444423119227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,7168,0.01796444422668881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,32,0.007000889215204451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,768,0.0057857777509424425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,3584,0.050888889365726046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,6144,0.08244888650046454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,6144,0.012444444828563266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,512,0.005097777893145879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,64,0.002985777747299936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,1536,0.02237777743074629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,65536,0.12221333715650771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,65536,1.3290559980604384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,128,0.002850666642189026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,64,0.0032960000551409195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,3584,0.15991200341118708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,3072,0.010268444816271463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,12288,0.06740088595284356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,16384,0.019360888335439894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,12288,2.6654417249891496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,256,0.005511999958091312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,3072,0.045181334018707275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,1536,0.02020088831583659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,1536,0.0058648888435628675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,6144,0.022264889544910852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,65536,0.20757155948215059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,3072,0.03463110989994473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,6144,0.04081777731577555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,128,0.004143110993835661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,1024,0.009379555781682333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,32,0.007787555456161499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,3072,0.010495999621020423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,64,0.002696000039577484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,1024,0.009309333231714036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,3072,0.020640888147883944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,2560,0.01219733307758967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,10240,0.13626044326358372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,12288,0.03860177927547031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,1536,0.0097884444726838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,128,0.007743111087216272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,512,0.0047564444442590075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,2048,0.006771555377377405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,1024,0.024878222081396315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,7168,0.08663466903898452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,6144,0.021013332737816706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,768,0.007289778027269576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,1536,0.06947022014194064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,768,0.005663110978073544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,2048,0.009568000005351173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,7168,0.012494222157531314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,3584,0.006432000133726332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,16384,2.776381386650933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,1536,0.011394666300879585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,16384,0.1598284509446886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,2560,0.2592159907023112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,10240,0.015260444747077094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,3072,0.014838222000333997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,2560,0.007277333074145847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,4096,0.008939555949634975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,1024,0.007544000115659501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,5120,0.1080720027287801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,2560,0.007727999654081132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,1536,0.02640266716480255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,65536,3.470831129286024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,32,0.038520889149771795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,768,0.006151111175616582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,768,0.006978666616810693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,1024,0.004950222041871813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,64,0.008388444781303406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,2560,0.011108444796668159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,2048,0.006011555592219035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,16384,0.026122666067547266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,10240,1.02619202931722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,128,0.009583999713261923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,6144,0.0143857780430052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,256,0.004400889078776042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,12288,0.044473777214686074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,10240,0.04630400074852837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,65536,0.06725155644946627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,2048,0.07937066422568427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,10240,0.03480266531308492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,4096,0.026958222190539043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,64,0.0026106666773557663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,3584,0.01483022173245748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,10240,0.08944177627563477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,4096,0.017493334081437852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,128,0.004593777573770947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,10240,2.23635376824273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,3584,0.012836444709036084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,256,0.007542221910423703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,16384,0.06150311231613159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,128,0.0051857779423395796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,8192,0.012141333685980903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,2560,0.046675556235843234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,5120,0.010283555421564314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,7168,0.05972355604171753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,8192,0.017736888594097562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,3584,0.012049777640236748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,256,0.0064426664676931165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,128,0.0053048887186580235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,7168,0.01223288890388277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,256,0.0074746666683091064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,32,0.008522666990756989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,16384,0.0730115572611491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,256,0.003648888733651903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,2560,0.017803554733594257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,1024,0.0063742221229606206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,65536,1.0976986355251737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,8192,0.01107377807299296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,64,0.00591111100382275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,512,0.0067528887755341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,2560,2.120063145955404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,10240,1.4774612850613063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,10240,0.008423999779754216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,16384,0.07520800166659884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,768,0.010262221925788457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,4096,0.033997333712048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,7168,0.04758933186531067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,256,0.006219555520349079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,6144,5.6930024888780375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,12288,0.01848266687658098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,2048,0.011572444604502784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,3584,0.1581102212270101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,7168,0.013549333645237817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,7168,0.009232000344329411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,6144,0.0099973330895106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,3072,0.053809778557883374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,128,0.009461333354314169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,6144,0.04446399874157376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,8192,0.018253333038753934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,10240,0.013515555196338229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,1536,0.004831999954250124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,2560,0.052368889252344765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,12288,0.01516266663869222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,256,0.009400000174840292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,3584,0.0075902218619982404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,512,0.004475555486149258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,4096,0.015562666787041558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,3584,0.00703022215101454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,4096,0.03860622313287523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,256,0.07176622417238024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,2048,0.0140364451540841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,768,0.0053671110007498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,65536,0.07566666603088379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,2048,0.04085422224468655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,12288,0.09340533283021714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,1024,0.011740444435013665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,8192,0.061625778675079346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,768,0.008802666432327693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,2560,0.08706577618916829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,512,0.0033173333439562055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,8192,0.013891556196742587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,2560,0.009840889109505547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,3072,0.007797333101431529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,65536,0.8317306836446127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,5120,0.01956266661485036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,2560,0.008551999926567078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,3072,0.01646399994691213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,8192,0.028592887851927016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,2560,0.006806222101052602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,1536,0.038422223594453596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,8192,0.016840888394249808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,10240,0.025207999679777358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,5120,0.14017332924736872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,2048,0.01884355478816562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,65536,0.12113955285814072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,768,0.005136888888147143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,128,0.002987555538614591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,3584,0.012895111408498553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,5120,0.01223288890388277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,256,0.0033200000309281876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,4096,0.01997599999109904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,65536,0.1827404499053955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,16384,0.11957421567704941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,768,0.013404444687896304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,1536,0.16269421577453613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,7168,0.09385244713889228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,8192,0.02850133180618286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,2048,0.012645333177513547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,128,0.0035208890007601846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,32,0.014551111393504672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,64,0.004848888764778773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,2048,0.012112889024946423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,256,0.004302222281694412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,3072,0.008088000118732452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,512,0.009845332966910468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,4096,0.017318222257826064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,2560,0.009744889206356471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,2560,0.08712000317043728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,128,0.032638221979141235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,2560,0.011329777538776398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,256,0.021710221966107685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,1536,0.021506667137145996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,64,0.005956444475385878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,3584,0.017853332890404593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,8192,0.3512720002068414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,7168,0.028500444359249536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,128,0.0028693332440323303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,2048,0.0069226668112807805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,512,0.00508355556262864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,32,0.0026168889469570587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,16384,1.1263511445787218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,10240,0.48696623908148867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,768,0.01795733306143019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,10240,0.008719999757077958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,12288,0.590810669793023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,1024,0.006707555717892117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,64,0.003076444483465619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,12288,0.26525243123372394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,128,0.006405333264006509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,64,0.012031111452314587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,256,0.005602666487296422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,6144,0.01016622202263938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,512,0.0034506666577524612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,10240,0.028096887800428603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,8192,0.012115555504957834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,768,0.008017777568764156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,512,0.00360000009338061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,65536,0.14465067121717665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,65536,0.07627555396821764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,12288,0.010267555713653564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,1024,0.008747555315494537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,1024,0.003952888978852166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,768,0.006112888869312074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,512,0.005527110977305307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,1024,0.005979555762476391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,32,0.0047066667013698155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,3584,0.019103111492262948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,128,0.005132444616821078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,4096,0.027427555786238775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,12288,0.017059556312031217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,1536,0.006476444502671559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,512,0.00906488878859414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,2048,0.006636444479227066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,128,0.005148444324731827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,4096,0.010055111514197456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,12288,0.014993778533405729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,5120,0.03149866726663377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,768,0.01794933279355367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,3584,0.03808355662557814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,256,0.0031902222997612427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,16384,0.055126223299238414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,65536,0.20480799674987793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,5120,0.041689776711993747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,128,0.002692444456948174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,65536,0.21234310997856987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,4096,0.01346577786737018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,512,0.003216000066863166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,10240,4.149468315972222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,12288,0.022239110536045496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,16384,0.27407288551330566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,6144,0.2034684419631958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,12288,0.015108444624476962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,256,0.0033964444365766314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,12288,0.01809955636660258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,2560,0.015399111641777886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,1536,0.008621333373917473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,5120,0.00904888908068339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,16384,0.2875857883029514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,3584,0.023822221491071913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,2048,0.00738844441043006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,32,0.014451555079884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,1536,0.005958222266700532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,7168,0.009483555952707926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,256,0.01779911087618934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,32,0.014298665854665967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,2560,0.01806311143769158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,3584,0.010306666294733683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,1536,0.006698666761318843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,5120,0.006212444355090459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,7168,0.011137777732478248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,3584,0.009143999881214565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,2560,0.03052177694108751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,32,0.00942488925324546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,1024,0.0058151111006736755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,5120,0.047543111774656505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,3584,0.011540444360838996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,2048,0.011669333610269757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,1536,0.005692444327804778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,1024,0.007300444775157505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,12288,0.015604444675975375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,256,0.004464000049564573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,512,0.004936000125275718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,6144,0.009581333233250512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,4096,0.07817866404851277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,16384,0.009648000200589498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,12288,0.016576889488432143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,3584,0.0090524446633127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,65536,0.08917688661151463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,12288,0.016862221890025668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,32,0.0028373334142896864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,8192,0.018038221531444125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,128,0.008265777594513362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,1536,0.011345778074529437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,12288,0.20526488622029623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,64,0.007160888777838812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,4096,0.2328693336910672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,16384,0.023219555616378784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,256,0.005422222117582957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,128,0.00407288885778851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,256,0.005582222094138463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,65536,0.678402688768175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,10240,0.024180443750487432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,3072,0.010818666881985135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,8192,0.012252444194422828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,3072,0.009942221972677443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,128,0.008785777621799046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,10240,0.18356711334652373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,4096,0.008979555633332994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,2560,0.007897777689827813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,2560,0.00630044440428416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,3584,0.008783111141787635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,8192,0.02671555512481266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,3072,0.010944888823562197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,3072,0.006364444477690591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,768,0.005902222047249476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,32,0.005167110926575131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,768,0.00543200017677413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,512,0.0034151110384199354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,2048,0.009992889232105678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,3072,0.006934222247865465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,16384,0.2765582137637668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,1536,0.00905511114332411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,2560,0.03512800070974562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,3072,0.013567110730542077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,4096,0.010564444793595208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,2048,0.005513777749405966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,3584,0.007673777639865875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,128,0.013637333280510373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,1536,0.009844444692134857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,12288,0.05384977658589681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,3072,0.021479111578729417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,1024,0.035424888134002686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,12288,0.03049244483311971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,65536,0.186063117451138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,2048,0.06692622105280559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,65536,1.1268417570326064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,512,0.007480888730949826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,10240,0.04201066825124952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,1024,0.00572533326016532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,7168,0.10386488835016887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,8192,0.02515999972820282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,7168,0.021753778060277302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,512,0.00555644432703654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,65536,0.10284711254967584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,12288,0.01806400054030948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,3584,0.015237333046065437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,1536,0.07411466704474555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,16384,0.008714666797055138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,64,0.002605333303411802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,16384,0.022406222091780767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,2560,0.020689777202076383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,12288,0.018385777870814007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,3584,0.01682044400109185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,12288,0.010478222535716163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,7168,0.025270221961869135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,5120,0.013432888521088494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,12288,0.032280001375410296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,32,0.004606222112973531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,1024,0.004969777746333016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,128,0.01824977828396691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,1024,0.004447111239035924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,1536,0.01629066632853614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,2048,0.005703111075692707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,65536,0.06857422325346205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,768,0.026208889153268602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,2048,0.007129777636792924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,4096,0.061121775044335254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,65536,0.14590666029188368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,1024,0.017076444294717576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,2048,0.007233777807818518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,8192,0.01775733298725552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,3072,0.009496889180607265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,3072,0.016238222519556682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,1024,0.0037831109431054858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,65536,0.08998933103349473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,1024,0.0038897775941424896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,4096,0.03314666615592109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,5120,0.46776713265313047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,1536,0.00453244439429707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,65536,0.08016444577111138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,65536,1.0351048575507271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,2560,0.007841777470376756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,128,0.003107555624511507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,2560,0.015987555185953777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,1536,0.00583022211988767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,256,0.004519111166397731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,2048,0.010277333358923594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,12288,0.06591733296712239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,10240,0.040510220660103693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,32,0.0035217776894569397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,12288,0.014161777165200977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,128,0.006474666711356904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,512,0.006798222247097228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,65536,0.06934844122992621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,10240,0.015232000086042615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,1536,0.04888622297181023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,10240,0.3804462220933702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,1024,0.012247111234400006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,16384,0.04931288957595825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,12288,0.07738488912582397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,1536,0.11620444721645778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,2560,0.008373333348168267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,256,0.004357333398527569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,512,0.0034151110384199354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,2048,0.0069395556218094295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,6144,0.012714666624863943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,256,0.005869333528810077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,12288,0.2508648766411675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,16384,0.07783644729190402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,10240,0.010104888843165504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,1536,0.007076444725195567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,3072,0.015435554915004306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,1024,0.0162471118900511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,2048,0.009506666825877296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,2560,0.009495110975371467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,512,0.021672000487645466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,256,0.007156444092591603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,16384,0.1890773375829061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,3584,0.006280000011126201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,64,0.004595555365085602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,5120,0.009222221871217092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,3584,0.031954665978749595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,65536,0.18644711706373426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,8192,0.03883822096718682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,7168,0.9844035042656792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,12288,0.08007644282446967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,7168,0.02755111124780443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,10240,0.020054222808943856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,1536,0.019119110372331407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,2048,0.008543999658690559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,512,0.048235555489857994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,6144,0.036107556687461004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,3072,0.09863911072413127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,256,0.007342221836249034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,256,0.0041582224269707995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,3584,0.02981511089536879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,512,0.039459556341171265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,128,0.0064382221963670515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,32,0.002612444468670421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,5120,0.009731555150614845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,3072,0.010983111129866706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,2048,0.013216888739003075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,768,0.003682666768630346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,6144,0.11616710821787517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,2560,0.012160888976520963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,6144,0.010255999863147736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,128,0.010195555786291758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,8192,0.05165955424308777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,8192,0.020741333564122517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,768,0.015758223003811307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,2048,0.005018666800525454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,128,0.00461422238085005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,10240,0.033879111210505165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,768,0.005853333406978183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,4096,0.014001778430408902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,32,0.03554666704601712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,128,0.0038044444388813446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,10240,0.08938133054309422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,2560,0.07890133062998454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,1024,0.017687999539905123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,256,0.004858666823969947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,7168,0.01611022154490153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,10240,0.3420426580641005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,65536,0.06992444727155897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,5120,0.02199733257293701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,512,0.007800888684060838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,4096,0.04677333434422811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,3072,0.007382222347789341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,1024,0.037598222494125366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,3584,0.014351111319330005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,1024,0.008447999755541483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,768,0.015222221612930298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,512,0.004365333252482944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,2048,0.007634667058785756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,5120,0.16449422306484646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,3584,0.01579466627703773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,1024,0.005684444473849402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,768,0.005736888696750005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,768,0.005113777601056629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,16384,0.017307556337780423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,4096,0.0181688881582684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,3584,0.010909332997269101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,768,0.013931555880440606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,1024,0.034584889809290566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,256,0.0052844443255000645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,10240,0.026760889424218073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,1536,0.03794222076733907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,3584,0.0103502223889033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,2048,0.011754666765530905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,8192,0.4364577664269342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,512,0.0034302222645945023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,7168,0.007728888756699032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,5120,0.03781955440839132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,2560,0.018783999813927543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,2560,0.007826666865083907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,5120,0.015440889530711703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,3072,0.10297866662343343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,2048,0.026557332939571802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,512,0.003334222154484855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,16384,0.04881777697139316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,2048,0.01277511070171992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,7168,0.4614195293850369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,512,0.005202666752868229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,1536,0.022828444838523865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,65536,0.4215822219848633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,64,0.002595555658141772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,128,0.013821333646774292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,8192,0.01293244461218516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,3584,0.009687999884287516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,16384,0.03771200113826328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,64,0.005592888842026393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,16384,0.09876533349355061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,2560,0.007425777614116669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,512,0.03900444507598877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,65536,0.28638845019870335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,7168,0.7864560021294488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,768,0.004815111143721474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,256,0.004690666579537922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,16384,0.8984177907307943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,16384,0.007694222033023834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,12288,0.02436977790461646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,2560,0.02943377693494161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,4096,0.021687999367713928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,128,0.003018666679660479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,2048,0.10508088933096991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,256,0.002891555635465516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,5120,0.009520000053776635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,3584,0.008408889174461365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,1024,0.006554666492674086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,64,0.008547555241319869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,6144,0.011050666371981302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,7168,0.01038577738735411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,16384,0.0203804439968533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,1536,0.004912888838185204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,16384,0.04861422379811605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,2048,0.0053288886944452924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,1024,0.00982844415638182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,64,0.03865422142876519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,10240,0.04607377780808342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,64,0.006879111131032308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,2048,0.053287999497519604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,2560,0.007071110937330458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,512,0.0053022222386466126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,1536,0.005596444424655702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,128,0.005072888814740711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,2560,0.13443822330898708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,1024,0.007270221908887227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,128,0.030809776650534734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,3072,0.205384890238444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,2560,0.015117333994971381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,7168,0.07226044601864286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,65536,0.07870399951934814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,64,0.005355555564165115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,16384,0.020561777883105807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,256,0.0064995553758409284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,128,0.007263999846246507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,128,0.0028035555862718155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,5120,0.011316444310877057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,8192,0.011723555624485016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,12288,0.019127110640207928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,5120,0.009977777798970541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,5120,0.01463644372092353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,12288,0.015399111641777886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,128,0.005530666559934616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,10240,0.7286275757683648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,2048,0.013297778036859302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,3584,0.013472889032628802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,10240,1.4775386386447484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,1024,0.004919111314747068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,1024,0.00998577806684706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,5120,0.029923554923799302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,10240,0.013849777479966482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,4096,0.018882667024930317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,3584,0.1279768943786621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,2048,0.006239111224810283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,128,0.00600000015563435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,4096,0.006440889090299606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,16384,0.07040266858206855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,2560,0.01917866700225406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,512,0.005449777675999536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,12288,0.2964417669508192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,8192,0.011353777514563667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,4096,0.008450667063395182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,2048,0.006665777828958299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,10240,0.037112000915739275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,65536,3.714113023546007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,768,0.006543111056089401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,4096,0.026330666409598455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,8192,0.02843733297453986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,6144,0.3417804506089952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,32,0.007263999846246507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,65536,0.22560622957017687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,10240,0.035968889792760216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,10240,0.02209866709179348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,6144,0.012253333297040729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,7168,1.4240781995985243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,1024,0.010132444401582083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,10240,0.04864799976348877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,2048,0.007121777368916406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,2560,0.008319999608728621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,64,0.03033422099219428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,65536,0.1333786646525065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,6144,0.02904799911710951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,3584,0.008991999758614434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,8192,0.025091555383470323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,1536,0.006341333190600078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,5120,0.008983111215962304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,2048,0.0053377776510185665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,128,0.0053119998839166425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,1536,0.006067555397748947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,10240,0.023044443792766992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,1536,0.0068764446510208984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,128,0.03727733426623874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,5120,0.013108444710572561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,2048,0.012439111040698158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,128,0.003128000017669466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,3072,0.08275910880830553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,256,1536,0.005706666658322017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,2048,0.006759111252095964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,65536,0.06678666671117146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,5120,0.04055288765165541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,32,0.003168888803985384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,12288,0.03551377852757772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,1536,0.07109066512849596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,3584,0.01310222182008955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,6144,0.024635555015669927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,512,0.005561777700980504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,3584,0.01479822231663598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,2048,0.00925066653225157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,2560,0.04576355550024244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,128,0.0028764444092909494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,65536,0.9798559612698026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,3584,0.35425421926710343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,4096,0.024530667397711012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,32,0.005463111317820019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,1024,0.009727111293209923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,8192,0.027790221903059218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,256,0.005412444472312927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,65536,0.0917102230919732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,768,0.011852444046073489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,768,0.007623111208279927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,128,0.005195555587609609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,256,0.0166631109184689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,65536,0.8561333550347222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,32,0.0032142222755485107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,2048,0.021560889151361253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,10240,0.01681422193845113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,1024,0.007128888534175024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,10240,0.006631111105283101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,12288,0.010820444259378644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,10240,0.006272888845867581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,6144,0.007440889047251807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,512,0.0053297777970631914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,3584,0.01723377737734053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,2048,0.007855999800893996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,512,0.004713777866628435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,7168,0.08857422404819065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,16384,0.9417395061916776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,16384,0.03251111176278856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,65536,0.21337689293755424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,5120,0.020062221421135795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,512,0.0064533332155810455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,65536,0.09662755330403645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,12288,0.16811645030975342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,1536,0.013626666532622443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,1024,0.006911110960774952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,2560,0.07756711377037896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,65536,0.11097777552074856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,8192,0.03477866782082452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,64,0.005070222334729301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,256,0.00425155543618732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,3584,0.011845333708657159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,3072,0.025251555773946974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,8192,0.012152888708644442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,12288,0.01754399968518151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,5120,0.012348444097571902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,10240,2.5663324991861978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,4096,0.04166311025619507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,32,0.0031679999083280563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,256,0.0070471109615431885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,6144,0.043065776427586876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,64,0.02121155626244015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,2560,0.047246221039030284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,3072,0.008895999855465358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,8192,0.02381688853104909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,5120,0.01771377854877048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,768,0.0038417776425679526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,3072,0.011485333244005838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,2048,0.034267554680506386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,3584,0.012834666503800286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,32,0.01777866648303138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,2048,0.041899555259280734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,1024,0.005104889058404499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,10240,0.018801778554916382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,512,0.0046453335218959385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,5120,0.012335999972290464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,12288,0.3702390988667806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,1024,0.00888000014755461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,256,0.004940444396601783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,768,0.009367110828558603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,6144,0.025080000360806782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,1536,0.008647110727098253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,256,0.0041848888827694785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,12288,0.0755795571539137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,768,0.00830577810605367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,5120,0.016749334004190233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,256,0.00795022232664956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,1536,0.035417778624428645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,5120,0.01646933290693495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,3584,0.046037332879172436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,10240,0.02036177780893114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,8192,0.014118222726715935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,2048,0.0737537807888455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,256,0.003100444459252887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,768,0.0037155557009908888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,2048,0.00868000007337994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,2048,0.006016888966162999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,2048,0.19355466630723742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,768,0.006649777707126405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,6144,0.11965155601501465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,256,0.00574933323595259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,5120,0.12487110826704238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,16384,0.044248888889948525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,256,0.006565333240562015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,1024,0.0070159998204973005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,3584,0.05445422066582573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,12288,0.014105778601434497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,2560,0.04832622077729967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,5120,0.006247111078765657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,12288,0.018423111902342904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,512,0.0032906666811969546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,6144,0.16897866461012098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,512,0.004925333377387789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,64,0.00254133323000537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,3072,0.006414222220579783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,4096,0.01891288823551602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,16384,0.006480888773997624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,3072,0.010703999963071613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,2560,0.014673777752452426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,32,0.0027146666414207886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,12288,0.008570666942331526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,4096,0.010102222363154093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,1024,0.007896888587209914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,3584,0.01629244453377194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,512,0.005712000032265981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,2048,0.006326222171386083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,6144,0.00959022260374493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,3072,0.00906133320596483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,3584,0.2262666755252414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,7168,0.015054222610261707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,3584,0.010713777608341642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,64,0.006299555715587404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,8192,0.03305244445800781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,65536,0.12033955256144206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,5120,0.09569866789711846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,16384,0.03461866577466329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,5120,0.008667555948098501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,32,0.004898666507667965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,2560,0.009920000202125972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,64,0.0025822222232818604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,3072,0.056707554393344455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,5120,0.015579554769727917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,32,0.008439111212889353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,2560,0.013041777743233575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,3072,0.00738666703303655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,512,0.11034488677978516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,12288,0.06832444667816162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,6144,0.012351999680201212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,512,0.00498488876554701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,5120,0.00925155563486947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,4096,0.024171556035677593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,512,0.004352000024583605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,1024,0.005364444520738389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,128,0.08447555700937907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,7168,0.012429333395428128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,7168,0.01053244454993142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,1024,0.005980444451173146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,65536,0.11632000075446235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,8192,0.015190222197108798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,5120,0.028528889020284016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,7168,0.13452711370256212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,2560,0.03129600154028998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,3072,0.006276444428496891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,1024,0.034725334909227155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,512,0.008424888882372115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,768,0.010315555665228102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,2048,0.008496888809733922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,128,0.0029075555503368378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,16384,0.03748622205522325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,2048,0.01963555481698778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,3072,0.01314222233162986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,32,0.006654222392373615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,10240,0.017616889543003507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,2048,0.19611822234259713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,8192,0.06037955813937717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,256,0.003157333367400699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,1024,0.004912000149488449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,2560,0.02091644373204973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,3584,0.01314933349688848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,2560,0.008022222254011366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,2560,0.00847111145655314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,3584,0.008719111482302347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,768,0.010567111273606619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,2048,0.012883555557992725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,10240,0.049795554743872754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,65536,0.22578133477105033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,768,0.005761777775155173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,3072,0.009145778086450364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,768,0.02134399943881565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,12288,0.5497164196438259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,3584,0.010299555957317352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,4096,0.012441777520709567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,3072,0.017090666625234816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,4096,0.014168888330459595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,768,0.03286044465170966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,768,0.08228444390826754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,12288,0.01980977753798167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,65536,0.24477688471476236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,2560,0.01722044414944119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,1024,0.026683555708991155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,8192,0.043734222650527954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,768,0.005191111316283544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,8192,0.018114666144053142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,2048,0.010964444114102257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,2048,0.1346648931503296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,12288,0.2216728925704956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,1536,0.0063386667105886675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,768,0.006047111004590988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,8192,0.13031821780734593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,5120,0.00904444439543618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,5120,0.013890667094124688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,256,0.005036444299750858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,64,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,3584,0.23733690049913195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,5120,0.03563466668128967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,1024,0.00424799985355801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,128,0.00573600000805325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,3072,0.015947555502255756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,64,0.022664889693260193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,10240,0.1608604457643297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,5120,0.1626746654510498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,768,0.008650666309727563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,32,0.006344888773229387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,10240,0.006892444358931647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,10240,0.019294222195943195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,2560,0.007329777710967594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,6144,0.012823999755912356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,1536,0.012149333126015134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,65536,0.07814755704667833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,256,0.005558222118351195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,4096,0.01053244454993142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,128,0.006743999818960826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,1024,0.007923555870850882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,2048,0.025370667378107708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,4096,0.007670222057236566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,6144,0.009501333038012186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,65536,0.14910044935014513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,10240,0.015726221932305228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,12288,0.05266044537226359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,12288,0.018411555223994784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,1024,0.00904177791542477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,3072,0.07053066624535455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,128,0.007749333149856991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,2560,0.15212621953752306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,1536,0.004690666579537922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,2560,0.028526223368114893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,8192,0.029714667134814795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,512,0.003660444584157732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,1024,0.24133422639634874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,10240,0.058529780970679395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,2048,0.007269333634111617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,4096,0.018972444865438674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,1024,0.06543644269307454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,2560,0.009977777798970541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,5120,0.006552888701359431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,64,0.005690666536490123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,64,0.007248000138335758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,12288,0.037943111525641546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,5120,0.04568355613284641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,768,0.005993777679072485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,7168,0.008079110748238033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,128,0.005127111242877113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,6144,0.05382400088840061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,512,0.0034391110142072043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,1024,0.009696889254781935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,1536,0.011869333684444427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,10240,0.02534399926662445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,512,0.005541333307822545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,4096,0.27564533551534015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,128,0.005610666755172942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,2560,0.05626222160127428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,65536,0.036631110641691424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,512,0.004794666750563515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,10240,0.01770666738351186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,512,0.013373333546850415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,4096,0.01718488832314809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,8192,0.01961599952644772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,2560,0.024812445044517517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,2048,0.009930666950013902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,2560,0.007777777810891469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,4096,0.01587022178702884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,256,0.0064275554484791225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,7168,1.3793857362535265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,8192,0.05474222368664212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,2560,0.0074311114019817775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,65536,0.0758204460144043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,12288,0.020282665888468426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,2048,0.02997244397799174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,65536,0.08109866910510592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,256,0.0041413332025210066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,64,0.004886222382386525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,32,0.003067555526892344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,512,0.009155555731720394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,3584,0.013273778061072031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,12288,0.04241866535610623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,16384,0.39681598875257706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,5120,0.013352000051074557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,5120,0.2789004378848606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,2048,0.006406222366624408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,4096,0.012920888761679331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,64,0.004836444639497333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,128,0.003346666693687439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,64,0.0037644443412621817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,768,0.008143111235565608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,4096,0.01692800058258904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,16384,0.0649991101688809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,16384,0.047148442930645414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,2560,0.017709333035680983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,5120,0.010581333604123859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,65536,0.03999377621544732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,512,0.003265777809752358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,768,0.003624888757864634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,2048,0.012216000093354119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,12288,0.018385777870814007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,3072,0.00811111099190182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,4096,0.016516443755891588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,4096,0.015512888630231222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,512,0.007716444631417592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,768,0.010777778095669217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,128,0.002891555635465516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,5120,0.009514666265911525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,2048,0.012615111139085559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,512,0.00979911122057173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,8192,0.025119110941886902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,768,0.003669333126809862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,7168,0.028004444307751123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,768,0.018360889620251127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,32,0.0052284445199701525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,768,0.008063111040327284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,3584,0.024999999337726172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,1024,0.011917333636018964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,10240,0.009119110802809397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,3584,0.016759110821617972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,128,0.005157333281305102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,768,0.00980888886584176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,256,0.008512000242869059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,2560,0.0103511114915212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,2048,0.008314666648705801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,1536,0.004724444614516364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,768,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,10240,0.01629155543115404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,256,0.011222222612963783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,512,0.005630222045713001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,1024,0.0058604445722368025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,3072,0.011957333319716983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,4096,0.10099999772177802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,768,0.04470488760206434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,65536,0.06779288583331637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,10240,0.01677600045998891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,7168,0.027453333139419556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,32,0.0036684444381131064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,7168,0.08454400300979614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,12288,0.027313777142100867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,5120,0.013021333350075616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,512,0.006715555571847492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,7168,0.0058355554938316345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,128,0.004952888935804367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,5120,0.023008000519540574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,12288,0.06523822413550483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,1024,0.006624889042642381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,256,0.0029813332690132987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,3584,0.014242667290899487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,8192,0.03368977705637614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,1024,0.0058088890380329555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,128,0.013367111484209696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,4096,0.008255999949243333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,128,0.006642666541867786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,256,0.005554666535721884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,1536,0.00453333349691497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,512,0.0046453335218959385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,128,0.007825777762466008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,2560,0.006039999839332368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,5120,0.015133332875039844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,64,0.0029279999434947968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,12288,0.12084355619218613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,4096,0.012698666916953193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,512,0.014829332629839579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,6144,0.0710808899667528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,2048,0.010294222169452244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,10240,0.014136888914638095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,32,0.002691555561290847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,7168,0.014762666490342883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,3584,0.01203555530971951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,6144,0.00833155545923445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,4096,0.019712888532214694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,10240,0.01808444493346744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,768,0.0064586665895250105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,10240,0.06168444289101494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,6144,0.010264889233642155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,128,0.003952888978852166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,4096,0.012973333398501078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,12288,0.025977777110205755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,1536,0.020720889170964558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,12288,0.046385778321160205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,512,0.0047786666287316214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,64,0.0031226666437255014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,64,0.0026728889594475427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,4096,0.008173333273993598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,16384,0.022445332672860887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,128,0.0042453333735466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,6144,0.010431110858917236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,4096,0.014484445254007975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,65536,3.094000074598524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,256,0.007048889166778988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,3072,0.0083155557513237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,128,0.009485333330101436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,4096,0.011050666371981302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,3072,0.014459555347760519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,32,0.006103111224042044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,256,0.003107555624511507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,256,0.005881777654091517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,5120,0.026017778449588354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,2560,0.47072532441880965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,3072,0.00775733341773351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,16384,0.06921066840489705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,256,0.00816000004609426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,32,0.005717333406209946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,7168,0.012873777912722694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,128,0.004231111043029361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,7168,0.028934223784340754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,5120,0.006959999601046245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,7168,0.07495822509129842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,2560,0.00702133360836241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,65536,0.08170933193630643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,3072,0.006320000108745363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,3584,0.007070221834712558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,7168,0.31740978029039174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,10240,0.44409065776401097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,65536,0.00924711094962226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,65536,0.1909724473953247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,6144,0.02181511123975118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,768,0.0037902221083641052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,768,0.008236444658703273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,512,0.010412444670995077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,1536,0.0064311110311084324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,7168,0.015210666590266757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,256,0.004313777718279096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,2560,0.012267555627557965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,8192,0.006696888970004187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,256,0.0036159998012913596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,2048,0.005906666732496685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,128,0.004068444586462445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,65536,0.3332151042090522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,12288,0.018557333283954196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,768,0.006636444479227066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,3072,0.006973333656787872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,2048,0.010653333531485664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,4096,0.01553244392077128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,3584,0.008580444587601556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,3584,0.015953777564896476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,32,0.024845333562956914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,4096,0.014013333453072442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,16384,0.03585777680079142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,32,0.015226667126019796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,1536,0.014117333624098035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,6144,0.011323555476135679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,65536,0.07169244686762492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,1024,0.009977777798970541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,512,0.005413333161009683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,1536,0.004658666749795278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,16384,0.022259554929203455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,2048,0.091940442721049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,64,0.004400889078776042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,768,0.005977777971161737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,6144,0.015671999918089975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,6144,0.008540444076061249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,128,0.00573955559068256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,512,0.0033857776886887024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,1024,0.008779555559158325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,8192,0.030710223648283217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,16384,0.05022755596372816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,7168,0.03288622366057502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,65536,0.08815911081102158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,512,0.007266666326257918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,16384,0.021091555555661518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,2048,0.0131217779384719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,1024,0.0040693332751592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,4096,0.018583999739752877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,1536,0.022138666775491502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,8192,0.040552000204722084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,4096,0.1440595520867242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,2048,0.006511111226346757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,3072,0.016189333465364244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,128,0.00544177782204416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,128,0.009720000127951304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,64,0.04237955477502611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,10240,0.17382310496436226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,7168,0.028004444307751123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,2560,0.012237333589129977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,1536,0.005656888915432825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,2048,0.012049777640236748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,3072,0.008192000289758047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,2560,0.007019555403126611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,5120,0.0063288890653186375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,65536,0.1193679968516032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,1536,0.06885955731074016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,7168,0.018400889303949144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,65536,0.5775404506259495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,1024,0.025262221693992615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,768,0.00442755553457472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,3072,0.011418666276666852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,12288,0.017089777522616915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,65536,0.07175199853049384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,3584,0.007348444726732042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,64,0.005214222189452913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,768,0.007306666837798224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,768,0.005491555564933353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,7168,0.019800888167487252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,1536,0.006125333408514659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,1536,0.004602666530344221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,128,0.006244444598754247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,768,0.0041351111398802865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,4096,0.016270221935378183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,2560,0.03738311264250014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,6144,0.006155555446942647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,10240,0.012162666353914471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,12288,0.018793778287039865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,4096,0.01701955497264862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,8192,0.012206222448084088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,256,0.006083555519580841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,128,0.009296889106432596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,1536,0.008457777400811514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,65536,0.3913875685797797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,2560,0.00776533368561003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,12288,0.014927110738224454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,2560,0.012558221817016602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,128,0.005172444300519096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,65536,2.6050542195638022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,64,0.006568000134494569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,1024,0.004911111046870549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,10240,0.03795911206139459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,2048,0.0052968888647026485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,8192,0.00600355573826366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,8192,0.017510222064124215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,10240,0.18176889419555664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,128,0.0035777779089079965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,3072,0.589570681254069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,768,0.004263111286693149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,16384,0.460119088490804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,768,0.08382311132219102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,16384,0.028520001305474177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,1024,0.018203554881943595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,6144,0.040595554643207125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,64,0.0031475555151700974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,4096,0.0904373327891032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,5120,0.029831111431121826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,5120,0.03127911024623447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,12288,0.3277333312564426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,2560,0.01143555591503779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,5120,0.03902755512131585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,768,0.0081404447555542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,8192,0.018050667312410142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,1536,0.006120000034570694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,6144,0.010214221974213919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,1024,0.021184000704023574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,12288,0.019183999962276883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,768,0.007693332930405934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,8192,0.0329386658138699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,2560,0.013537777794731988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,10240,0.01811733345190684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,3584,0.007875555919276344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,1024,0.004487110922733943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,5120,0.021925333473417494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,768,0.004284444368547863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,5120,0.027952889601389568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,5120,0.023711111810472276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,128,0.004676444249020683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,2560,0.014797333214018079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,10240,0.1114799976348877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,16384,0.20671733220418295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,64,0.0027520000520679685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,32,0.008752000000741746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,3072,0.00656088896923595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,5120,0.03781244489881728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,3584,0.01090311093462838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,3584,0.04088711076312595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,5120,0.039965334865782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,65536,0.3426995542314317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,65536,0.7874995337592231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,256,0.0035217776894569397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,16384,0.034033778640959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,8192,0.15922755665249294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,1536,0.006072000082996156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,16384,0.40363645553588867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,768,0.004280000097221798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,1024,0.005774222314357758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,16384,0.2886035442352295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,10240,0.006264888991912206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,6144,0.006333333336644703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,2560,0.13460177845425075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,3584,0.009310222334331935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,2048,0.007984889050324758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,3584,0.013569778038395775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,1024,0.003959111041492886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,6144,0.02366311185889774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,256,0.003277333246337043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,7168,0.024719999896155462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,16384,0.05140533381038242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,256,0.006326222171386083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,65536,0.28379466798570424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,2560,0.027569777435726587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,2560,0.006583111153708563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,512,0.004211555752489301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,8192,0.35644178920321995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,32,0.013594666288958656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,768,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,2560,0.08375377787484063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,3072,0.2872968779669868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,2048,0.0075937774446275495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,7168,0.006506666541099548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,256,0.0047564444442590075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,16384,0.09052977959314983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,32,0.0029164445069101122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,2560,0.0063528890411059065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,5120,0.023342221975326538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,3584,0.07934666342205472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,7168,0.024687111377716064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,4096,0.011004444625642566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,1024,0.004134222037262387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,65536,0.06495822138256498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,65536,0.07309866613811918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,3072,0.011167111496130625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,7168,0.022623111804326374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,1024,0.028010666370391846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,128,0.004964444372389051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,128,0.005772444523043103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,2560,0.006568888823191325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,2048,0.006791111081838608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,6144,0.026566222310066223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,64,0.005612444546487596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,12288,0.022633777724372015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,6144,0.010133333504199982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,5120,0.016173332929611206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,4096,0.03751644492149353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,5120,1.4520710839165583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,32,0.007844444778230455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,65536,0.9522231419881185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,12288,0.21616178088717988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,6144,0.020090666082170274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,32,0.003132444495956103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,7168,0.006198222024573221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,2048,0.02126400007141961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,4096,0.0355333321624332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,2560,0.28876177469889325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,16384,0.044421331750022046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,32,0.007272000114123027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,1024,0.007747555772463481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,2048,0.007636444436179266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,64,0.0040995557275083335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,768,0.007024889190991719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,32,0.01072088877360026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,10240,0.13932622803582087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,768,0.017084444562594097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,4096,0.008223111430803934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,7168,0.013765333427323235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,256,0.006894222150246303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,1536,0.006264888991912206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,2560,0.05652533637152778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,8192,0.08361955483754475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,2560,0.402599122789171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,2048,0.009708444277445475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,6144,0.060538669427235924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,32,0.03690133492151896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,32,0.006897777732875612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,64,0.01201600001917945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,16384,0.04398311177889506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,3584,0.06224799818462796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,512,0.004632888982693355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,1536,0.012162666353914471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,1536,0.00978488889005449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,128,0.005032888717121548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,2560,0.35061245494418675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,256,0.01682133310370975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,12288,0.0329386658138699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,10240,0.4306577841440837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,10240,0.49382935629950625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,2048,0.00757244477669398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,1536,0.0111191107167138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,2048,0.1866977744632297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,65536,0.4323982132805719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,2560,0.010674667027261523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,4096,0.0451182226339976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,2560,0.021879111727078755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,1536,0.004856889032655292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,3072,0.017587555779351127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,8192,0.059528887271881104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,6144,0.018631999691327412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,768,0.024304888314670984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,65536,0.2453493277231852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,1024,0.007909333540333642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,65536,0.09839555290010239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,256,0.019148444135983784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,6144,0.020250666472646926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,7168,0.02018666598531935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,65536,0.10302311182022095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,10240,0.028396444188223943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,256,0.0036355555057525635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,65536,0.506132443745931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,1024,0.005320000151793162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,5120,0.248764435450236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,2560,0.029230223761664495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,1536,0.04246133234765795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,6144,0.026569777064853247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,3584,0.028578665521409776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,256,0.002974222310715251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,768,0.009453333086437648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,7168,0.03793155484729343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,4096,0.7352079815334744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,10240,0.013379555609491138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,10240,0.03502666619088914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,10240,0.023919110496838886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,12288,0.022431999444961548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,6144,0.029733333322736952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,32,0.0024968888610601425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,6144,0.07624444696638319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,512,0.00564533347884814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,1536,0.006452444526884291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,512,0.0070462218589252895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,3584,0.00701422244310379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,128,0.006263999889294307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,12288,0.018062222335073683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,2048,0.014367111855083041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,32,0.03573333223660787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,3584,0.013008000122176277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,512,0.0033093332830402586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,3072,0.01625866691271464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,256,0.012139555480745105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,6144,0.005772444523043103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,2560,0.020593777298927307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,1024,0.006225777582989798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,7168,0.022355554832352534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,12288,0.015551111764378019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,65536,0.12164266904195149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,3072,0.02849422229660882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,4096,0.013537777794731988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,10240,0.014315555493036905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,3072,0.05637066894107395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,512,0.005630222045713001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,1536,0.5747191111246744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,128,0.0034097778714365433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,64,0.0030951110853089225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,3584,0.012173333101802401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,128,0.0244195560614268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,6144,0.023453333311610754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,7168,0.011763555308183035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,12288,0.024998222788174946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,128,0.009209777745935652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,7168,0.010808000134097205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,3584,0.046943998999065824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,16384,0.040538665321138166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,768,0.005657777604129579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,768,0.01088444474670622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,12288,0.33235110176934135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,2560,0.007695111135641734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,256,0.00443111111720403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,8192,0.10069688823488022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,3584,0.01163733336660597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,3584,0.2754862308502197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,12288,0.017863111363516915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,10240,0.017874666386180453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,4096,0.011794666449228922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,512,0.005902222047249476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,2048,0.010984000232484607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,768,0.008398222426573435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,8192,0.010788444843557147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,64,0.0026462222966882917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,2560,0.023800000548362732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,1536,0.03280177712440491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,1536,0.010090666512648264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,10240,0.1524968942006429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,6144,0.11550754970974392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,12288,0.018277333842383493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,32,0.002508444504605399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,3072,0.009441778063774109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,3072,0.10443911287519668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,7168,0.008617777791288164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,768,0.01798577772246467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,64,0.005703111075692707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,32,0.005365333209435145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,7168,0.025384000606007043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,3584,0.06846311357286242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,3072,0.058861335118611656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,3584,0.07888622416390313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,8192,0.02179555594921112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,16384,0.06571200158860949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,5120,2.0675716400146484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,16384,0.13300711578792995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,3584,0.01035377797153261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,4096,0.009081777599122789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,1536,0.006806222101052602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,1536,0.005511111269394557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,256,0.008051555189821456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,1536,0.005049777941571342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,4096,0.31813867886861164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,2048,0.016314667132165697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,3072,0.007840000092983246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,5120,0.018485332528750103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,1536,0.011946666571829053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,1536,0.004657777647177379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,8192,0.01882133384545644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,1024,0.006229333165619109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,32,0.005231110999981563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,64,0.0053102220926019884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,3584,0.00905688852071762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,2048,0.006089777582221561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,32,0.002630222174856398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,16384,0.027912888262006972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,6144,0.23979287677341035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,32,0.007276444799370236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,8192,0.01977599991692437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,2048,0.05009955498907301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,65536,0.2364257706536187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,4096,0.006467555546098285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,12288,0.05718666977352566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,1024,0.0169840008020401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,10240,0.0401777790652381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,12288,0.018954666124449838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,3072,0.006443555570311016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,16384,0.01293777757220798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,7168,0.016912889149453905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,3584,0.07428266604741414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,10240,0.21598045031229654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,8192,0.01993333299954732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,256,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,12288,0.2095724476708306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,5120,0.04708710975117154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,2048,0.16843288474612764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,5120,0.008749333520730337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,10240,0.03160266743765937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,1536,0.004958222309748332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,10240,0.029847999413808186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,256,0.003112888791494899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,6144,0.014105778601434497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,16384,0.031712889671325684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,1536,0.00738666703303655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,2560,0.009263111485375298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,65536,0.10775199863645767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,12288,0.04100799891683791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,2560,0.00865955568022198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,1536,0.007515555454625024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,1024,0.004761777818202972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,6144,0.11584799819522434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,1024,0.006244444598754247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,1536,0.01071733319097095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,7168,0.0064035554726918536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,10240,0.31634312205844456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,10240,0.04004089037577311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,256,0.005379555539952383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,10240,0.02825333343611823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,3584,0.010679999987284342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,8192,0.0746320022477044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,256,0.0030586665703190696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,7168,0.01681511104106903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,16384,0.3429440127478705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,16384,0.05795288748211331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,65536,0.07620533307393391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,768,0.00832622249921163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,10240,0.015079110860824585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,65536,0.06880799929300944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,128,0.008889777792824639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,256,0.01827822294500139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,4096,0.011005333728260465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,64,0.007065777977307637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,7168,0.727583090464274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,10240,0.012228444218635559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,16384,0.023230221536424425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,2560,0.007256000406212277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,128,0.006250666661394968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,16384,0.016958221793174744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,8192,0.013421333498424955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,1024,0.005664000080691443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,16384,0.0188400000333786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,12288,0.02108888824780782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,768,0.0036071112586392295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,32,0.0030373332815037835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,5120,0.8018088870578342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,2560,0.07815644476148817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,4096,0.018555555078718398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,8192,0.016906667086813185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,2048,0.008330666356616551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,6144,0.11232444975111221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,16384,0.022525333695941504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,1024,0.005994666781690385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,6144,0.13449511263105604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,16384,0.023746665981080797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,6144,0.012988444831636218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,16384,0.1937662230597602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,3584,0.015997333659066092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,6144,0.010975110861990186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,4096,0.0069306666652361555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,3072,0.016846223009957206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,3072,0.0074675555030504865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,1536,0.01015733347998725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,1536,0.021514667405022517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,2048,0.014312000738249885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,1024,0.008627555436558193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,1536,0.006937777830494775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,32,0.0027537778433826235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,7168,0.01399466726515028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,3584,0.029263112280103896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,3072,0.015110222829712762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,512,0.07973866992526583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,256,0.022708444131745234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,256,0.008356444537639618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,256,0.026749332745869953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,10240,0.17806843916575113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,64,0.003519999898142285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,3072,0.014700444208251106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,32,0.0031831111345026228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,768,0.005473777651786804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,65536,0.3413813379075792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,256,0.00388355553150177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,3072,0.014854222536087036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,4096,0.010819555156760745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,6144,0.01591999994383918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,256,0.005230222311284807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,12288,0.5665689044528538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,3072,0.012452444268597497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,256,0.006682666639486949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,64,0.003116444374124209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,512,0.00835111074977451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,1024,0.0058248887459437055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,32,0.008185777399275038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,2048,0.012685333689053854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,3072,0.007416889071464539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,12288,0.041936000188191734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,4096,0.008073777788215214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,2048,0.00702044450574451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,16384,0.024719999896155462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,7168,0.12616177399953207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,7168,0.007059555914666917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,1024,0.004871110949251386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,8192,0.09296355644861858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,4096,0.007745777567227681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,5120,0.006309333360857434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,128,0.0028159999185138275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,256,0.010260444548394945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,16384,0.040752000278896756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,3072,0.04337955514589945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,128,0.006380444599522485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,256,0.004857777721352047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,2560,0.09764622317420112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,2048,0.016207999653286405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,12288,0.08388266960779826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,1536,0.007400000260935889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,32,0.008839999636014303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,6144,0.015831111205948722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,8192,0.011261333194043903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,3072,0.007140444384680853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,8192,0.25771644380357533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,3584,0.0084915558497111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,64,0.0028035555862718155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,6144,0.028039998478359644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,4096,0.01867999964290195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,768,0.00499288903342353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,256,0.00318755561278926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,5120,0.00944800012641483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,128,0.018439110782411363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,32,0.0025395556456512874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,2560,0.02657777733272976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,1024,0.005759999983840519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,1536,0.006007110906971826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,2048,0.011666666302416058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,3584,0.23997865782843697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,1536,0.009000000026490953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,3072,0.017198221551047433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,6144,0.02500711050298479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,6144,0.021566222111384075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,256,0.0031519999934567344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,12288,0.23996088239881727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,5120,0.011032889286677042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,1024,0.010045333041085137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,10240,0.05376355515586006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,1536,0.008461332983440822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,32,0.007631111476156447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,1024,0.003899555653333664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,16384,0.05113155643145243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,6144,0.009957333405812582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,256,0.004320888883537716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,2560,0.019131556153297424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,256,0.003069333318206999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,768,0.0170160002178616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,32,0.014775110615624322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,2048,0.013757333159446716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,128,0.0026791110220882627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,7168,0.018736888964970905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,2560,0.045444445477591626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,2560,0.012684444586435953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,2560,0.009275555610656738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,768,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,2560,0.009452444811662039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,10240,0.017453332742055256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,16384,3.4090364244249134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,6144,0.021902221772405837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,65536,0.06605777475568983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,6144,0.011386666860845355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,10240,0.013648000028398303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,5120,0.02276266614596049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,2048,0.00942488925324546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1536,4096,0.008244444098737504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,12288,0.11133600605858697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,64,0.0028968888024489083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,8192,0.011746666497654386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,1536,0.15184799830118814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,16384,0.07092889149983723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,3584,0.01016355554262797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,3584,0.019333332777023315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,65536,0.07040177451239692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,4096,0.2272906568315294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,2048,0.008343111309740279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,8192,0.006313777632183499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,10240,0.4078693389892578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,768,0.003834666477309333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,256,0.03394577900568644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,1536,0.012228444218635559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,256,0.005223111146026188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,1536,0.01978222197956509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,8192,0.03243466549449497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,16384,0.011608888705571493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,3072,0.018352000249756705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,1536,0.012227555943859948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,7168,0.012883555557992725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,3584,0.006534222099516127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,65536,0.5492346551683214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,768,0.005192000004980299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,3072,0.017508443858888414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,16384,0.013811555173661975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,768,0.7710417641533746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,16384,0.05526222123040093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,3072,0.01760444376203749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,128,0.0038133333954546186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,3584,0.06907111406326294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,3072,0.00906133320596483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,512,0.009319999979601966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,8192,0.016599110431141324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,3072,0.007669332954618666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,256,0.004504000147183736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,2048,0.12149511443244086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,12288,0.16053244802686903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,512,0.005961777849329843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,5120,2560,0.00981955561372969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,2048,0.011339555184046427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,1024,0.007071110937330458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,3072,0.008663111262851292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,8192,0.013145777914259167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,1024,0.09358488851123387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,3584,0.007366221812036302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,10240,0.024896888269318476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,7168,0.02807377775510152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,7168,0.04994933472739326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,2048,0.01089422239197625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,65536,3.3054186503092446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,2560,0.009696889254781935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,10240,0.02125599980354309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,5120,0.012707555459605323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,7168,0.03300266795688205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,2560,0.010840888652536603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,4096,0.008882666627566019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,64,0.0031831111345026228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,65536,0.5089253319634331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,12288,0.024273778001467388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,256,0.013595555391576556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,8192,0.8510648939344617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,768,0.1459804508421156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,12288,0.0595102243953281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,256,0.006705777926577463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,1024,0.006739555547634761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,3072,0.0075902218619982404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,2560,0.014557333456145393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,4096,0.016807110773192514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,128,0.0038159998754660287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,1024,0.007361777954631382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,1024,0.008357333640257517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,3584,0.011858666936556498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,1536,0.010144000252087912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,128,0.0027502222607533136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,16384,0.010607110957304636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,2048,0.0058853332367208265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,256,0.018735110759735107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,3584,0.06456977791256376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,5120,0.01180266671710544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,65536,0.39635467529296875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,64,0.01607111096382141
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,16384,0.008987555901209513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,65536,0.07420711384879218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,10240,0.014749333262443542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,768,0.019024888674418133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,8192,0.03065955638885498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,3584,0.018992889258596633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,10240,0.04399289025200737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,12288,0.018623110320832994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,5120,0.025052444802390203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,3584,0.014505777094099255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,7168,0.030249777767393324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,3072,0.005987555616431766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,4096,0.03933066791958279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,256,0.003012444410059187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,65536,0.3689582347869873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,2048,0.021360889077186584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,12288,0.3740124437544081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,1024,0.02144977781507704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,65536,0.24291822645399305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,6144,0.028636445601781208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,1024,0.006305777778228124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,256,0.005222222043408289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,5120,0.02281155520015293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,512,0.005107555538415909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,3072,0.015764445066452026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,128,0.006575999988449945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,2560,0.01017066670788659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,2560,0.019119999474949308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,6144,0.0109120003051228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,12288,0.028986665937635634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,8192,0.01163733336660597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,64,0.002814222127199173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,512,0.005618666609128316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,512,0.009700444837411245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,1024,0.012278222375445895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,256,0.005673777725961473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,16384,0.17431199550628662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,4096,0.15296622117360434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,3072,0.011658666862381829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,3072,0.019129777948061626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,1536,0.030838221311569214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,5120,0.013026666310098437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,10240,0.1068097750345866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,2560,0.017683555682500202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,512,0.006862222320503659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,5120,0.010321777727868822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,3584,0.026239110363854304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,32,0.004598222259018156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,1536,0.004423999951945411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,5120,0.011970666547616323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,1536,0.006597333484225803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,1536,0.02926666537920634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,256,0.03574577636188931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,6144,0.01593777702914344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,16384,0.05473777651786804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,128,0.018593778212865193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,3584,0.021740444832377966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,32,0.05042666527960035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,7168,0.00637511122557852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,65536,0.0775431129667494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,3072,0.013825777504179211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,65536,0.1285573376549615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,32,0.006883555402358373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,2560,0.011384888655609555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,1024,0.00658044425977601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,8192,0.023597333166334365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,256,0.052319112751219005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,2560,0.047523554828431874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,768,0.03062044580777486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,128,0.005919111271699269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,512,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,1024,0.005474666754404704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,5120,0.01110577748881446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,3584,0.06325955523384942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,8192,0.00647288892004225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,8192,0.03245688809288873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,10240,0.2610231240590413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,32,0.005362666729423735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,12288,0.042430222034454346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,4096,0.011391999820868174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,5120,0.016647110382715862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,3072,0.00981155534585317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,512,0.014003554979960123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,7168,0.22307822439405653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,512,0.005143110950787862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,3072,0.04390311241149902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,12288,0.020093333390023973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,16384,0.020536889632542927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,3072,0.45844088660346133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,1536,0.006204444501135085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,64,0.008115555677149031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,512,0.006600889066855113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,3584,0.009474666582213508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,6144,0.02958222230275472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,4096,0.0210515558719635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,6144,0.013755555782053204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,256,0.006762666834725275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,4096,0.012739555703269111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,12288,0.016927111479971144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,4096,0.007937777373525832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,768,0.005730666634109285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,6144,0.030091553926467896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,128,0.006772444479995304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,2560,0.013493333425786761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,8192,0.07611022392908733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,256,0.003949333396222857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,32,0.0025439999169773525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,10240,0.030856887499491375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,7168,0.1985928879843818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,65536,0.18145599630143908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,16384,0.03281689021322463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,3584,0.027765333652496338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,8192,0.011717333561844297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,768,0.05993688768810696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,1024,0.005108444641033809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,3072,0.007108444141017065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,6144,0.011789333489206104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,4096,0.008415111237102086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,8192,0.012849777936935425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,256,0.0041164445380369825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,16384,0.12597155570983887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,64,0.006697777658700943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,16384,0.024515555964575872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,2560,0.23878399531046549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,3584,0.009728888670603434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,512,0.0034924443397257063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,768,0.005055111315515307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,12288,0.016022221909628976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,64,0.0032231110251612137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,256,0.008260444634490544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,2560,0.007060444189442529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,8192,0.007735110819339752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,1536,0.005668444352017508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,128,0.004234666625658671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,8192,0.40287110540601945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,64,0.006492444624503453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,10240,0.027990221977233887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,1024,0.005629333357016246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,64,0.014059555199411182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,2048,0.007840889195601145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,1024,0.006331555545330048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,8192,0.018183999591403537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,4096,0.014832889040311178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,7168,0.0539644459883372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,3072,0.013171555267439948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,3584,0.018345778187115986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,256,0.00490399988161193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,4096,0.40273155106438535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,3072,0.013095999757448832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,16384,0.17989067236582437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,4096,0.03164533442921109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,10240,0.08195644617080688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,1536,0.02622755534119076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,4096,0.007820444802443186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,3072,0.026268444127506677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,7168,0.04203822215398153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,2560,0.1734764443503486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,768,0.10298933585484822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,10240,0.06753155257966784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,768,0.004616888860861461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,8192,0.02179733415444692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,7168,0.3075573444366455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,768,0.005220444252093633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,12288,0.0348328881793552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,7168,0.030606223477257624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,4096,0.01331822243001726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,12288,0.20956267250908744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,2048,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,5120,0.008751110898123847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,6144,0.0459591117170122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,6144,0.009484444227483537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,12288,0.04124177826775445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,2560,0.0536133348941803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,16384,0.011704888608720569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,2560,0.007318221860461765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,256,0.0042160000238153665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,128,0.008801778157552084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,32,0.005397333453098933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,32,0.002543111021320025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,256,0.03657333387268914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,512,0.004550222307443619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,7168,0.029343111647499934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,10240,0.04184800055291918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,1024,0.013382222089502545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,512,0.005968889014588461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,2560,0.04482488830884298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,2048,0.01052977806992001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,128,0.005063111169470681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,10240,0.06089422437879774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,1536,0.04377777708901299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,12288,0.11007466581132676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,6144,0.08403111166424221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,3072,0.01479022204875946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,3584,0.009889778163697984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,1024,0.004534222185611725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,32,0.014522666732470194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,3584,0.007391110890441471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,512,0.01482222146458096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,256,0.006816888848940532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,6144,0.02223555578125848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,1536,0.005520888914664586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,65536,0.7897262043423123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,12288,0.05198666784498426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,256,0.20430844359927705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,6144,0.011200889117187925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,3072,0.015272888872358533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,3072,0.025399999486075506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,3072,0.008968888885445064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,8192,0.03263466556866964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,2048,0.13847200075785318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,2048,0.007109333243634965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,65536,0.3526613447401259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,7168,0.028833776712417603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,5120,0.01201866649919086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,16384,0.2149262163374159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,5120,0.21224621931711832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,32,0.0031351111829280853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,128,0.004517333375083076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,6144,0.02384000023206075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,2560,0.006753777878151999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,128,0.003375111147761345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,65536,0.1826559967464871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,65536,0.18022667037116158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,16384,0.01311822235584259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,7168,0.04030311107635498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,10240,0.22543022367689344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,3584,0.010608889162540436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,512,0.15108356210920545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,7168,0.017697778013017442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,5120,0.04227644536230299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,768,0.005097777893145879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,64,0.005511999958091312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,65536,0.012624888784355588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,65536,0.0927342242664761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,6144,0.026269333230124578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,32,0.00674311113026407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,6144,0.048639999495612256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,256,0.039239999320771955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,3072,0.013667555318938361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,32,0.00868000007337994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,10240,1.4654817581176758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,65536,0.27557510799831814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,12288,0.04273066586918301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,512,0.005005333158704969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,64,0.0058853332367208265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,65536,0.543631977505154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,3072,5.053582933213976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,8192,0.02311733365058899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,768,0.016866667403115165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,3072,0.012959111067983838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,8192,0.013032000097963544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,1536,0.005927111125654644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,3584,0.013545778062608508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,16384,0.021167111065652635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,12288,0.02481866710715824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,256,0.005642666584915585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,65536,14.880162556966146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,65536,0.31599201096428764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,6144,0.010866666833559671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,12288,0.21352977222866484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,1536,0.010178666975763109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,1536,0.014538667268223233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,12288,0.019243554936514962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,6144,0.023359111613697473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,2560,0.04562133219507006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,3584,0.016978666186332703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,32,0.014049778381983439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,12288,0.011191111471917896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,128,0.005151999907361137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,8192,0.11544266011979844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,512,0.004779555731349521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,128,0.0039022221333450745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,1024,0.005486222190989389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,1536,0.050092445479498975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,512,0.004810666872395409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,8192,0.03999377621544732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,2560,0.007718222008811102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,8192,0.015331554743978711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,64,0.006421333385838403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,32,0.005494222044944763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,10240,0.0210479994614919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,1024,0.004265777766704559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,3072,0.01258399999803967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,64,0.0028986665937635633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,512,0.0032684444967243406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,2560,0.02219999995496538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,16384,0.022925333844290838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,256,0.009865778187910715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,4096,0.01776088939772712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,16384,0.022237333986494277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,3584,0.047500444783104785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,256,0.006837333242098491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,1024,0.005947555518812603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,768,0.034972445832358465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,6144,0.08335555262035793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,12288,0.2075048949983385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,2048,0.006702222343948152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,64,0.0031475555151700974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,128,0.005104889058404499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,128,0.007096889118353526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,32,0.005277333160241445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,5120,0.008416888614495596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,64,0.0026106666773557663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,10240,0.023388443721665278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,16384,6.546411726209853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,3072,0.009544889132181803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,7168,0.014119999276267158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,6144,0.017086222767829895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,1024,0.10668800274531047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,2048,0.008085333638721043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,1024,0.008432000047630733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,1536,0.00637511122557852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,5120,0.017295999659432303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,1536,0.006698666761318843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,16384,0.021847999758190576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,32,0.004250666747490565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,512,0.006484444356626934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,12288,0.17437422275543213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,128,0.004687110996908612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,10240,0.07854311333762275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,8192,0.09650044308768378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,32,0.004831999954250124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,256,5120,0.0075235557225015424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,1536,0.010640000303586325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,5120,0.015434667468070984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,10240,0.04202666547563341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,256,0.013173333472675748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,1024,0.017238222890430026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,12288,0.024736000431908503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,10240,0.048657778236601085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,512,0.004809777769777509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,32,0.004754666652944353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,3584,0.007941332956155142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,768,0.006072000082996156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,64,0.005128889034191768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,2048,0.01070666644308302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,4096,0.07335733042822944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,3072,0.007619555625650618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,8192,0.02829066581196255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,512,0.005053333524200651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,6144,0.011156444748242697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,512,0.01974311139848497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,6144,0.020095111595259774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,3072,0.026779555612140234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,16384,0.026213333010673523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,64,0.005442666510740916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,3584,0.007852444218264686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,32,0.0030702222138643265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,5120,0.02120888895458645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,1536,0.011319110790888468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,64,0.016650666793187458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,128,0.005047111047638787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,2048,0.010871999793582492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,128,0.01664977769056956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,3584,0.028347555134031508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,1536,0.011380444798204633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,7168,0.015840889679061044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,128,0.004218666503826777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,128,0.005517333332035277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,65536,0.23817600144280326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,1536,0.0053635554181204895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,128,0.004435555388530095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,2560,0.05393866697947184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,32,0.0068417779273457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,1024,0.034669333034091525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,128,0.003939555750952827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,512,0.011687999798191918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,12288,0.015624889069133334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,256,0.0033893332713180115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,12288,0.02878755662176344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,2560,0.006065777606434292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,4096,0.006595555692911148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,64,0.006842666616042455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,1024,0.006360888895061281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,7168,0.015791111522250705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,5120,0.01533511115445031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,3072,0.01704355577627818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,3072,0.006615110983451207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,1536,0.10831022262573242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,8192,0.04553955462243822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,2048,0.013156444662147097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,1536,0.009881777895821465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,7168,0.06041600306828817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,3584,0.016859556237856548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,2048,0.028483556376563177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,1536,0.2942799992031521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,5120,0.014649778604507446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,10240,0.10031288862228394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,1024,0.06811022096210055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,512,0.007249777515729268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,1024,0.027951998843087092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,3584,0.010303110712104374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,512,0.019075555933846366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,256,0.003972444269392225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,2560,0.022691556149058874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,7168,0.018944889307022095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,1024,0.026491555902693007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,32,0.00637155564294921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,7168,0.006578666468461354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,768,0.041286223464541964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,5120,0.021244444780879553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,1024,0.012139555480745105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,6144,0.10840000046624078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,65536,0.04553866717550489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,1536,0.0124195557501581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,2048,0.005954666684071223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,3072,0.007701333198282454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,12288,0.02146133283774058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,10240,0.028280887338850234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,4096,0.011586666935020022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,2560,0.016158221496476066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,32,0.003007111036115222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,256,0.04680800106790331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,256,0.004089777668317159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,1536,0.018008000320858426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,16384,0.7396684222751193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,2048,0.011184888581434885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,8192,0.03137777911292182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,6144,0.27900976604885525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,65536,0.21661067008972168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,512,0.0054560001525614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,128,0.013439999686347114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,128,0.0053582220441765254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,4096,0.3615102238125271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,3584,0.009734222458468543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,768,0.004019555532270008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,256,0.004570666700601578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,2560,0.006791111081838608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,5120,0.02664799988269806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,12288,0.20989955796135795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,256,0.005417777846256892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,256,0.005160000175237656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,1024,0.0041662222809261745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,2048,0.016568000117937725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,6144,0.11744533644782172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,65536,0.2727208932240804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,512,0.005747555444637935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,4096,256,0.004987555659479565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,3584,0.00739555557568868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,5120,0.04021599888801575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,6144,0.017631111873520743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,1024,0.0058711109062035876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,5120,0.02163822286658817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,12288,0.021876444419225056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,2048,0.01144444445768992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,2048,0.013943110903104147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,32,0.006268444574541516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,64,0.006616888774765863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,4096,0.020486222373114694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,65536,0.06613244613011678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,2560,0.009452444811662039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,2560,0.006991110742092133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,3584,0.009168888959619734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,16384,6.932448069254558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,6144,0.01627377834584978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,12288,0.049919111861122974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,65536,7.673960367838542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,10240,0.02176711128817664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,16384,0.1459911134507921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,256,0.02407022151682112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,2048,0.007339555356237624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,4096,0.010296000374688042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,2560,0.006070222291681502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,4096,0.011603555745548673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,3584,0.1926186614566379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,3584,0.015770667129092746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,7168,0.5849608845180935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,1024,0.00397155558069547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,3072,0.03529333406024509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,1024,0.0047333331571684945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,256,0.0031519999934567344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,2560,0.08804533216688369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,2048,0.011376000112957425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,5120,0.03073599934577942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,512,0.0057795556883017225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,8192,0.032969776127073504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,768,0.009256000320116678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,2560,0.006771555377377405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,12288,0.028912001185946997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,10240,0.017729777428838942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,7168,0.015420445137553744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,8192,0.020601777566803824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,768,0.01865333318710327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,2560,0.01092088884777493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,4096,0.02085599965519375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,6144,0.02516533268822564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,2560,0.009878222313192155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,5120,0.013904000322024027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,2560,0.008909333083364699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,768,0.006937777830494775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,2048,0.01385155568520228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,6144,0.013047999805874295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,5120,0.021140444609853957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,6144,0.009540444446934594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,3072,2.616101370917426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,65536,0.18089600404103598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,6144,0.012716444830099741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,7168,0.22042844030592176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,3584,0.011771555576059552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,3072,0.006846222198671765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,2560,0.028396444188223943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,16384,0.13164088461134169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,32,0.005231110999981563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,4096,0.7899066607157389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,7168,0.04767555660671658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,3072,0.012655999925401477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,1024,0.009351111120647853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,32,0.003564444267087512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,7168,0.01457155578666263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,5120,0.011916444533401065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,768,0.0046471113132105935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,65536,0.2615786658393012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,3072,0.04540266593297323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,1536,0.005522666705979242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,7168,0.02234933276971181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,6144,0.013388444152143268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,256,0.01260444439119763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,12288,0.015571556157535978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,5120,0.011648889217111798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,256,0.00692622239391009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,2048,0.005250666704442766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,128,0.005954666684071223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,8192,0.1485733323627048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,5120,0.016759999924235873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,768,0.020017777880032856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,5120,0.015525332755512662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,6144,0.01314133322901196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,6144,0.20108266671498617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,512,0.003875555677546395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,12288,0.018197332819302876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,768,0.003843555433882607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,128,0.006663110935025745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,5120,0.019899555378490023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,2048,0.006181333214044571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,65536,1.5960249370998805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,1536,0.00555377784702513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,8192,0.016037333342764113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,768,0.008536000218656328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,3584,0.32820532057020396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,3584,0.009117333425415887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,2560,0.014420444766680399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,768,0.005633777628342311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,1024,0.010259555445777046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,64,0.004840888910823398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,32,0.0032506665835777917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,128,0.029653333955340918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,1024,0.0041653331783082755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,6144,0.014675555957688225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,65536,0.1409777800242106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,7168,0.03313333458370633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,1536,0.010121777653694153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,128,0.005936000082227919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,64,0.0037786666717794207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,3072,0.014347554908858405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,8192,0.014157333307796054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,512,0.003568888952334722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,32,0.0068959999415609576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,256,0.005375999957323074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,256,0.0029377777957253983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,1024,0.007661333514584436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,12288,0.01776266594727834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,1536,0.01201333353916804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,6144,0.12070488929748535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,512,0.009676444033781687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,8192,0.01275022245115704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,7168,0.021065778202480737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,512,0.004873777843183941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,1024,0.005240889059172736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,1536,0.014655111564530266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,1536,0.006992888947327931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,3584,0.022794667217466567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,6144,0.015557333827018738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,5120,0.025591111845440332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,5120,0.018012444178263348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,3584,0.02693155573474036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,256,0.007458666960398356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,2048,0.016343999240133498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,128,0.005504888792832692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,3072,0.01949155496226417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,7168,0.09464177820417617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,2560,0.015577778220176697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,256,0.004927999857399198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,512,0.0052897776994440295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,5120,0.00976533359951443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,256,0.005712888720962737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,256,0.005655111124118169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,256,0.010519111322032081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,5120,0.01834844383928511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,6144,0.014434667097197639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,1024,0.010024000373151567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,7168,0.01681777834892273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,2048,0.011058666639857821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,8192,0.031261334816614784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,512,0.005687110953860813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,2560,0.006803555621041193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,10240,0.03894755575391982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,256,0.015085332923465304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,3584,0.011180444724029966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,1536,0.007358222372002072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,8192,0.07424266470803155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,768,0.006222222414281633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,10240,0.009272000028027428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,7168,0.02485688858562046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,2560,0.12472000386979844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,256,0.016042666302786935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,10240,0.09777866469489203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,3584,0.008468444148699442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,64,0.0026862221873468826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,512,0.005433777968088786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,12288,0.0424542228380839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,64,0.0026417778184016547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,64,0.005109333329730564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,6144,0.015963556038008798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,1536,0.04714399907324049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,1024,0.005872888697518243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,1024,0.005461333526505365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,7168,0.021554667088720534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,512,0.036659555302725896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,32,0.005782222168313132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,2048,0.041202667686674334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,64,0.0057653333577844836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,32,0.02128977742460039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,768,0.0038648889296584656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,10240,0.05589510997136434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,12288,0.21962755256228975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,1024,0.010371555884679159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,32,0.007139555282062954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,256,0.0030862221287356485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,2560,0.027978665298885767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,3072,0.018548443913459778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,16384,0.02386933399571313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,6144,0.0934879978497823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,5120,0.02648266653219859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,3072,0.013069333301650153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,2560,0.00777688870827357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,7168,0.038811554511388145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,3072,0.031197332673602637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,5120,0.022408000297016565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,5120,0.030396444929970637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,128,0.005202666752868229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,16384,0.010576888918876648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,8192,0.44831201765272355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,2560,0.25280711385938853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,6144,0.0875120030509101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,6144,0.07568444146050347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,512,0.0047182221379545005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,8192,0.04981599913703071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,7168,0.2324248949686686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,4096,0.010031999813185798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,5120,0.01723288827472263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,1024,0.022487110561794702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,5120,0.02542488939232296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,768,0.01125511113140318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,256,0.012528888881206512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,12288,0.07328889105055067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,3072,0.0074675555030504865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,128,0.004504888835880491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,12288,0.045864890019098915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,5120,0.017056889004177518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,12288,0.5252328978644477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,7168,0.03467288944456313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,4096,0.010466666685210334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,768,0.01181244436237547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,3072,0.1636364459991455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,128,0.0034764446318149567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,10240,0.018914666440751817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,10240,0.33390577634175617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,1536,0.02163377735349867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,4096,0.008303110798199972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,5120,0.5914737913343642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,7168,0.1794133318795098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,5120,0.011389333340856763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,32,0.0041457778877682155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,256,0.008436444732877944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,32,0.010052444206343757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,32,0.0029564443975687027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,5120,0.006774222271309958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,16384,0.022681777675946552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,10240,0.10976977480782403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,65536,0.24041599697536895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,5120,0.014119111829333834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,12288,0.04696177774005466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,16384,0.008593777815500895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,512,0.005774222314357758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,1536,0.004628444297446145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,5120,0.008708444734414419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,7168,1.0998409059312608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,3072,0.06215200159284803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,5120,0.010334221853150262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,4096,0.015114666687117683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,2048,0.008656889200210571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,512,0.00574311117331187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,12288,0.1589111089706421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,32,0.0047351109484831495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,1536,0.02241155505180359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,64,0.00282577777074443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,32,0.007732444339328342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,768,0.009672000176376766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,32,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,256,0.002871111035346985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,3072,0.008607111043400234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,64,0.002591111179855135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,16384,0.013839999834696451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,1024,0.0041333333485656315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,1536,0.0063902222447925145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,512,0.006769777586062749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,12288,0.04244000050756666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,12288,0.016116445263226826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,2048,0.015059555570284525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,65536,2.421812481350369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,6144,0.02924799919128418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,65536,0.8546142048305936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,768,0.006562666760550604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,2048,0.010808000134097205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,512,0.027034666803148057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,65536,0.1797999938329061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,512,0.019768888751665752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,3584,0.025469332933425903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,1536,0.00509688879052798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,128,0.005635555419656966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,1024,0.00675644435816341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,4096,0.01871555546919505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,2048,0.01221866657336553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,4096,0.008185777399275038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,64,0.0026835555003749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,7168,0.0787680016623603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,1024,0.005993777679072485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,7168,0.0277173337009218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,2560,0.007168889045715332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,5120,0.022855111294322546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,7168,65536,0.1378364430533515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,128,0.0037004442678557504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,64,0.002581333327624533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,65536,0.07763199673758613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,256,0.005160000175237656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,512,0.02479822271400028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,8192,0.006593777901596493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,2560,0.011994666523403592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,4096,0.00925244473748737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,2048,0.005722666780153911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,3584,0.008952889177534316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,8192,0.012650666965378655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,768,0.0037724446091387006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,32,0.008571555217107138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,8192,0.10985333389706081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,768,0.012245333029164208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,3584,0.009530666801664565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,3072,0.006862222320503659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,32,0.0027466666781240036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,32,0.00272088891102208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,65536,0.1276871098412408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,512,0.004391111019584867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,1024,0.025031111306614343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,3072,0.009434666898515489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,256,0.005187555733654234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,4096,0.021634666456116572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,3584,0.06986488898595174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,8192,0.013987556099891663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,4096,0.011532444092962476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,10240,0.06687733199861315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,32,0.0029804443733559716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,12288,0.03252355588807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,768,0.00906488878859414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,3584,0.008665777742862701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,3584,0.09865599870681763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,65536,0.20593955781724718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,3584,0.02640266716480255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,512,0.014959999256663851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,3584,0.01404088901148902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,12288,1.8452782101101344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,1024,0.0069982219073507525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,1024,0.0074106670088238185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,10240,0.5123164388868544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,6144,0.011403555671374003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,1024,0.0047217777205838096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,768,0.009591999981138442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,4096,0.14113599724239773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,256,0.0052728888889153796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,128,0.02180355621708764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,2560,0.011559999651379056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,6144,0.024118221468395654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,5120,0.0503715541627672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,6144,0.01758666667673323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,16384,0.09925155507193671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,32,0.006421333385838403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,2560,0.006790222393141852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,3584,0.020640000700950623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,3072,0.01036088913679123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,4096,0.01941244469748603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,1024,0.004977777600288391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,768,0.01297688898113039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,10240,0.03772177630000644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,16384,0.051607112089792885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,128,0.0027866665687825945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,2560,0.0061919999619325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,10240,0.030983110268910725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,768,0.008373333348168267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,256,0.005129777722888523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,128,0.005361777626805835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,3584,0.06879733006159465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,6144,0.009915555516878763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,1024,0.0048942222363419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,65536,0.10941333240932888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,32,0.011869333684444427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,1024,0.007675555845101674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,12288,0.041309333509869046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,10240,0.04414933257632785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,16384,0.007095110913117726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,128,0.0026728889594475427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,7168,0.012079999678664737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,10240,0.05828444163004557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,2048,0.05024977856212192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,6144,0.04222133424546984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,65536,0.12337156136830647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,768,0.005399999933110342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,10240,0.04070933328734504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,10240,0.015815110670195687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,32,0.0068711112770769335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,12288,0.048826667997572154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,6144,0.12574399842156306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,2560,0.006169777777459886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,3072,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,3072,0.008329778081840938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,4096,0.08747200171152751
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,3584,0.007916444705592262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,2560,0.012326222327020435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,12288,0.028930667373869155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,3584,0.0207306659883923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,2048,0.007089777953094906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,3584,0.007224000162548489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,5120,0.012236444486512078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,3072,0.011230222052998014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,256,0.005024888863166173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,3584,0.018128888474570382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,4096,0.13320088386535645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,512,0.005004444470008214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,16384,0.631147543589274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,3072,0.07104355759090848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,12288,0.04452800088458591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,7168,0.010245333115259806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,1536,0.008560889297061497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,3584,0.024145777026812237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,16384,0.02529866662290361
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,12288,0.16283288266923693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,3072,0.008043555749787224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,6144,0.17070489459567598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,1024,0.004156444635656145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,65536,0.03652888867590163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,32,0.03196533189879523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,512,0.0037866665257347953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,64,0.005201777650250329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,768,0.01959733333852556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,3584,0.010642666783597736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,3072,0.01593955523437924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,3584,0.007970666719807519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,65536,2.595874574449327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,512,0.004607111215591431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,16384,0.029984000656339858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,2560,0.012813333008024426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,3072,0.060940444469451904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,128,0.0037288889288902283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,256,0.00656088896923595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,256,0.013581333061059317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,16384,0.055204444461398654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,12288,1.7970765431722004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,1536,0.012591111163298288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,12288,0.03892533315552606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,3584,0.030437333716286555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,2048,0.29854488372802734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,65536,0.20267111725277373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,64,0.08803733189900716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,4096,0.013944889108339945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,5120,0.03647289011213515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,768,0.013312000367376538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,512,0.0038746665749284956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,256,0.005756444401211209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,10240,0.01460800071557363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,1024,0.01110400011142095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,8192,0.26838310559590656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,128,0.004570666700601578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,512,0.011472889118724398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,1024,0.004588444613748127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,7168,0.2367937829759386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,512,0.01257955531279246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,65536,0.04166933231883579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,5120,0.006585777633719974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,128,0.002834666727317704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,3072,0.02409244411521488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,512,0.005472888963090048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,8192,0.06959555546442668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,16384,0.01015733347998725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,10240,0.21072622140248617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,2048,0.02479822271400028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,1536,0.005879111174080107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,16384,0.0698888897895813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,2560,0.007443555527263218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,8192,0.3061750994788276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,128,0.004881777697139316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,512,0.0053591111467944244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,5120,0.469504886203342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,256,0.003432000055909157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,4096,0.06077422036064995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,512,0.0148026661740409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,4096,0.011513777905040316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,768,0.028008000718222723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,5120,0.011375111010339526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,12288,0.06502577993604872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,64,0.004980444494220945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,3584,0.01645866698688931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,1024,0.003931555483076307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,512,0.007595555649863348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,4096,0.11179021994272868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,1536,0.005380444642570283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,10240,0.019116444720162284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,1024,0.005950222412745158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,128,0.006125333408514659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,768,0.010644444160991244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,32,0.002946666752298673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,2048,0.007099555598364935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,768,0.0038195554580953387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,2560,0.011531555818186866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,32,0.009014222357008193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,12288,0.01574577722284529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,12288,0.05224622289339701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,6144,0.04054311248991225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,6144,0.008206222620275285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,128,0.0028764444092909494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,5120,0.03112799922625224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,5120,0.09244088994132148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,64,0.0033528889632887314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,10240,0.08535289102130467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,6144,0.019201777047581144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,768,0.009339555270142024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,2560,0.007595555649863348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,10240,0.015056888262430826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,4096,0.005962666538026597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,10240,0.01329422245422999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,65536,0.06977155473497179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,16384,0.0064195555945237475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,65536,0.06779822376039293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,65536,0.6206337610880533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,7168,0.016142222616407607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,5120,0.021750221649805706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,3584,0.007954667011896769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,512,0.4331599871317546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,16384,0.06247378057903714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,3072,0.23190577824910483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,2560,0.018623110320832994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,16384,0.006646222124497096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,12288,0.02050755586889055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,16384,0.039275556802749634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,128,0.005414222263627582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,3584,0.021137777302000258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,8192,0.022259554929203455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,3072,0.010455999937322406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,12288,0.024327110913064744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,8192,0.028231110837724473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,5120,0.012704888979593912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,10240,0.026348445150587294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,7168,0.012596444123321109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,12288,0.04852177699406942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,128,0.005488889084921942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,64,0.0058106668293476105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,7168,0.012723555167516073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,4096,0.010103999740547603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,256,0.004030222280157937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,7168,0.06601778003904554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,4096,1.7483510971069336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,5120,0.04891022377543979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,64,0.008058666355080074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,1536,0.006808888994985157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,1024,0.013507555756303998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,12288,0.6550666491190592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,7168,0.014009777042600842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,16384,0.07304000192218356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,1536,0.0067528887755341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,8192,0.14214577939775255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,2560,0.01148888882663515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,64,0.003944888710975647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,1536,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,2048,0.0069395556218094295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,128,0.003621333175235324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,2560,0.006552888701359431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,7168,0.020206222931543987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,128,0.006220444622966979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,2048,0.010597333312034607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,16384,0.022655111220147874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,12288,0.02127200033929613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,16384,0.5895039770338271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,6144,0.015853333804342482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,7168,0.013640888863139682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,768,0.006238222122192383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,5120,0.006239999913507038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,12288,0.15294578340318468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,2560,0.005946666830115848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,12288,0.08289511336220635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,12288,0.2603742281595866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,32,0.0028284444577164124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,1024,0.01314133322901196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,16384,0.03157866663402981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,3584,0.013080000049538083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,5120,0.6957564353942871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,512,0.010278222461541494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,64,0.0029306666304667792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,4096,0.041908442974090576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,1536,0.006549333532651265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,7168,0.01147911118136512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,10240,0.033654222885767617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,65536,0.05395288930998909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,5120,0.14524977737002903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,8192,0.03292622168858846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,65536,0.20392266909281412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,1536,0.05589155356089274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,768,0.009200000100665623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,1024,0.00498933345079422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,6144,0.019745777050654095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,32,0.0033119999700122406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,2048,0.030155556069480047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,256,0.005224888937340842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,65536,0.2300168938106961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,16384,0.02117777698569828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,1024,0.011439111497667102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,10240,0.0074648890230390764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,4096,0.032935112714767456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,6144,0.026312000221676294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,65536,0.007862221863534715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,7168,0.02514044443766276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,1536,0.010597333312034607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,12288,0.016844444804721408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,64,0.0053546664615472155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,256,0.00980266680320104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,1024,0.0069306666652361555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,1024,0.00600888869828648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,6144,0.011220444407727985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,8192,0.03264888789918687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,3072,0.01555288831392924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,512,0.05467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,1024,0.007138667007287343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,10240,0.03394133183691237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,2048,0.06955911053551568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,768,0.007917332980367873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,3072,0.0069146665434042616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,16384,0.07176533010270861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,7168,0.031488888793521456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,1536,0.007069333559936947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,512,0.005391110976537068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,1024,0.038564443588256836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,4096,0.019241778386963736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,256,0.003080000066094928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,7168,0.0277804434299469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,256,0.013343999783198038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,768,0.008008889026112026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,10240,0.07856622007158068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,65536,1.1106711493598092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,64,0.004680888934267892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,16384,0.03788444399833679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,64,0.009047110875447592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,1024,0.004420444369316101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,65536,0.15359911653730604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,12288,0.06088088618384468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,5120,0.03882933325237698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,64,0.005205333232879639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,16384,0.33958133061726886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,16384,0.028943998946083918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,3584,0.007223111059930589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,2048,0.006018666757477655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,8192,0.07880177762773302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,128,0.0053048887186580235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,256,0.003718222181002299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,12288,0.01742844449149238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,7168,0.016843555702103507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,1536,0.0054551110499435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,12288,0.08372000190946792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,1024,0.009638222555319468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,8192,0.05122311247719658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,256,0.006812444577614467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,1536,0.006165333506133821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,65536,0.28999021318223744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,5120,0.012019555601808759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,3072,0.006494222415818109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,8192,0.05584177705976698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,64,0.0052044445441828836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,7168,0.06012711260053846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,64,0.005636444522274866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,4096,0.01518311103185018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,256,0.005613333235184352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,6144,0.26118487781948513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,512,0.008050666915045844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,7168,0.01642933322323693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,3072,0.01872711049185859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,128,0.005049777941571342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,3072,0.03205777870284186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,3072,0.30227910147772896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,256,0.006504000061088138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,8192,0.23119465510050455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,3584,0.0277164445983039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,8192,0.07182844479878743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,6144,0.019554666346973844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,65536,1.0635449091593425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,8192,0.06286933024724324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,512,3072,0.008055999875068665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,16384,2.0446205139160156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,768,0.00740622232357661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,7168,0.03277511066860623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,2048,0.0053671110007498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,128,0.03180266751183404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,32,0.0068604445291890045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,2560,0.007805333369308048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,1024,0.0037475555307335327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,2560,0.0064222220745351576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,256,0.005410666680998272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,32,0.005341333233647876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,128,0.006738666858938005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,8192,0.03096533483929104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,3584,0.017539555827776592
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,1536,0.011745777395036487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,1024,0.0057848890622456866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,7168,0.31348000632392037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,8192,0.02260444396071964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,256,0.005091555416584015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,6144,0.04682844546106127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,7168,0.013747555514176687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,3072,0.03084711233774821
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,4096,0.27360444598727757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,7168,0.015799111790127225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,65536,0.10310400194591945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,3072,0.01755466726091173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,128,0.008811555802822113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,12288,0.01869422197341919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,3584,0.08040889104207356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,256,0.005690666536490123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,128,0.002741333304180039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,768,0.006383111079533895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,65536,2.9871323903401694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,512,0.06748621993594699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,1536,0.013884444203641681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,2048,0.0845679971906874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,3072,0.009659555223253038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,64,0.0074142225914531285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,2560,0.009167110754383935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,10240,0.09795466396543716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,5120,0.023181334137916565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,7168,0.27067910300360787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,2048,0.00655377780397733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,1024,0.0057555557125144536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,6144,0.009451555709044138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,1536,0.01017599966790941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,6144,0.015054222610261707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,64,0.007144889069928064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,64,0.002566222308410539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,5120,0.014316444595654806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,3072,0.09018222490946452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,6144,0.014383110735151501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,6144,0.013101333545313941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,2560,0.0070017774899800616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,768,0.04569333460595873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,4096,0.013845333622561561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,3072,0.041371554136276245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,10240,0.01478488908873664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,128,0.004955555415815777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,128,0.0026977778308921387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,2560,0.010973333484596677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,65536,0.24426311916775176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,12288,0.033837334977255926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,16384,0.03046311272515191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,512,0.010449777874681683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,128,0.0027928888383838865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,3584,0.009186666872766282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,3072,0.025148444705539282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,12288,0.08669955862892999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,64,0.002642666714058982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,2560,0.00700444479783376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,256,0.004543111142185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,16384,0.3600346777174208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,10240,0.03398488958676656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,8192,0.015233778291278414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,3072,0.011798222031858234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,512,0.005985777825117111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,16384,0.06305244233873156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,1024,0.007088888850477006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,1024,0.04278577698601616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,65536,0.07890399959352282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,12288,0.006072000082996156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,2560,0.015407111909654407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,2048,0.008799110849698385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,4096,0.010510221951537661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,3584,0.013224889006879596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,2560,0.01035733355416192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,16384,0.08430666393703884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,128,0.0031226666437255014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,12288,0.07313511106703016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,512,0.005477333234416113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,512,0.003399111123548614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,512,0.004061333421203825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,256,0.029447111818525527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,7168,0.011128000087208219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,65536,0.04572355416085985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,6144,0.012447111308574677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,32,0.005780444376998478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,2048,0.012005333271291522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,128,0.00519466648499171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,10240,0.019866666860050626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,64,0.0025457777082920074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,1536,0.011281777587201862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,1536,0.02985066506597731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,3584,0.006991110742092133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,10240,0.07960355281829834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,512,0.024538666009902954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,3584,0.008091555701361762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,1024,0.005032888717121548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,768,0.007816000117195977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,512,0.020192889703644645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,12288,0.018415111634466384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,768,0.00535733335547977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,2560,0.018569777409235638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,16384,0.4099760055541992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,5120,0.012431999875439538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,512,0.006528888725572162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,1024,0.013229333692126803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,10240,0.03345333205329047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,10240,0.01811555524667104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,12288,0.0418835547235277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,8192,0.053780443138546415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,512,0.003653333418899112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,4096,0.007821333077218797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,2048,0.01423911088042789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,6144,0.00906488878859414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,256,0.005551110953092575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,12288,0.02604622145493825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,256,0.1347688833872477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,512,0.018538667096032035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,3072,0.010006222460005019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,2560,0.01897333396805657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,768,0.02125511070092519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,1024,0.005670222143332164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,1024,0.00638488887084855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,10240,0.014525334040323893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,2048,0.006745777610275481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,256,0.0032151111712058387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,7168,0.014208889669842191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,1024,0.005437333136796951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,768,0.00582311095462905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,2048,0.03489244315359328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,10240,0.028551999065611098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,4096,0.024265777733590867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,128,0.00601333338353369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,65536,0.6864764425489637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,256,0.015555555621782938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,10240,0.04246844516860115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,7168,0.02028888960679372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,6144,0.010710222025712332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,4096,0.011307555768224927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,256,0.005503111001518037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,32,0.005416000054942236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,1024,0.06521777974234687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,2560,0.015603555573357476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,512,0.0041991112132867174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,5120,0.020593777298927307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,12288,0.019116444720162284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,2560,0.31673065821329754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,6144,0.007544889218277401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,2048,0.022788445154825848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,256,0.004902222090297275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,512,0.06157066424687704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,3072,0.011355555719799466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,2560,0.010059555371602377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,6144,0.012422222230169507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,16384,0.02239199976126353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,256,0.00424888895617591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,512,0.005688888745175467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,1024,0.003887999802827835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,3584,0.014720888601409065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,2560,0.00778399987353219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,65536,0.0729902254210578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,2560,0.007662222617202335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,6144,0.017709333035680983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,6144,0.016241777274343703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,1536,0.007967111137178209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,5120,0.022271111607551575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,10240,0.026327111654811438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,512,0.0034604445099830627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,2560,0.005685333162546158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,256,0.004657777647177379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,512,0.00608711110221015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,16384,0.03977510995335049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,1024,0.007420444654093848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,1024,0.008249777886602614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,4096,0.00908799966176351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,768,0.006181333214044571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,65536,0.2562977737850613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,1024,0.026714666022194758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,7168,0.01641244524055057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,128,0.004455111092991299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,3584,0.01035199976629681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,8192,0.019271110494931538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,128,0.004221333397759331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,10240,8.69439951578776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,3584,0.01292799992693795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,256,0.004808889081080754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,6144,0.01183733344078064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,5120,0.02000266644689772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,3072,0.02628533376587762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,2048,0.0074826669361856245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,16384,0.05612711111704508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,10240,0.006704888823959563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,7168,0.012120889292822944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,4096,0.008368888662921058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,7168,0.027727999620967444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,6144,0.03615466753641764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,1024,0.0063662222690052455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,3584,0.007205333146784041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,2048,0.0064293332397937775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,768,0.00721244431204266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,16384,0.07228000296486749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,10240,0.12694399886661106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,768,0.01012711144155926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,2560,0.03682933251063029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,10240,0.46910932328965926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,5120,0.021683555510309007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,64,0.0029306666304667792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,5120,0.009484444227483537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,3072,0.025556445121765137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,2560,0.48233864042494035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,4096,0.007066667079925537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,2560,0.07804889149136014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,65536,1.3152782652113173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,32,0.002959111084540685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,3072,0.018872888551818002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,8192,0.010120888551076254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,32,0.0034542222403817703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,6144,0.10225600004196167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,768,0.004724444614516364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,5120,0.1894684500164456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,2048,0.005344888816277186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,3072,0.016623111234770883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,768,0.014509333504570855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,16384,2.042252434624566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,32,0.003630222131808599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,128,0.005388444496525659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,16384,0.6519466506110297
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,256,0.0035964445107513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,8192,0.015311110350820752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,7168,0.013785777820481194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,4096,0.011004444625642566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,3072,0.01680266691578759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,256,0.0059262220230367445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,12288,0.1633395618862576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,4096,0.03821688890457153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,5120,0.008223999705579545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,256,0.0047271110945277745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,10240,0.018193778064515855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,16384,0.06171111265818278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,10240,0.01463733282354143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,4096,0.007925333248244392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,768,0.008764444126023186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,8192,0.5207609070671929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,3072,0.034225778447257146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,256,0.005152889009979037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,12288,0.12331555949317084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,1024,0.011911999848153857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,4096,0.014001778430408902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,128,0.0029297777348094513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,256,0.0047253333032131195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,512,0.006718222051858902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,1024,0.00526311124364535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,16384,0.04813955558670891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,1024,0.005442666510740916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,4096,0.2016835610071818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,128,0.006294222341643439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,32,0.003310222178697586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,16384,0.00792266676823298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,1024,0.006911110960774952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,256,0.004287999951177173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,1536,0.007181333170996771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,256,0.005342222336265776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,12288,0.5131431155734593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,5120,0.10596177975336711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,16384,0.037992000579833984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,4096,3.520583894517687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,32,0.007038222418891058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,6144,0.020589333441522386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,4096,0.012990222209029727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,6144,0.022408000297016565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,2560,0.007947555846638149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,256,0.004480000171396467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,65536,0.9208933512369791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,12288,0.02913155489497715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,1024,0.011873777541849347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,256,0.013205333716339536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,6144,0.015536889433860779
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,1024,0.007374222079912822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,1536,0.01425066590309143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,1024,0.008780444661776224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,6144,0.015434667468070984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,256,0.005447111195988125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,3584,0.012076444096035428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,32,0.007347555624114142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,2048,0.009922666682137383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,512,0.008801778157552084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,5120,0.24812176492479113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,2560,0.007400000260935889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,16384,0.013542222479979197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,65536,0.10981955793168809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,16384,0.04032355546951294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,3584,0.01051111105415556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,512,0.0033866665843460294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,128,0.22091733084784615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,7168,0.013225778109497495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,4096,0.01070844464831882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,256,0.011590222517649332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,3072,0.0063680000603199005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,10240,0.01553600033124288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,16384,0.28270043267144096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,12288,0.021866665946112737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,1536,0.012721777790122561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,4096,0.00816000004609426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,2560,0.011704888608720569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,16384,0.17895821730295816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,10240,0.007034666836261749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,1536,0.02959111001756456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,16384,0.07901955313152738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,2560,0.013902222116788229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,16384,0.04416355490684509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,4096,0.13954133457607693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,7168,0.02084622283776601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,1536,0.0047111109726958806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,3072,0.04788088798522949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,6144,0.023387556274731953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,512,0.0189057770702574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,256,0.005316444569163852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,3584,0.10131466388702393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,32,0.010880889164076911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,32,0.016844444804721408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,5120,0.03133599956830343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,4096,0.01161688897344801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,7168,0.02428177826934391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,1024,0.007065777977307637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,256,0.05309955610169304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,1536,0.009806222385830348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,256,0.005586666779385672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,6144,0.09443733427259658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,7168,0.012551999754375882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,16384,0.08311466375986735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,7168,0.5341688791910807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,12288,0.7102026409573026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,768,0.00665244460105896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,1536,0.00812355594502555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,4096,0.0118568887313207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,32,0.0024604443460702896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,5120,0.02731200059254964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,16384,0.030199110507965088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,256,0.0028080000645584534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,10240,0.0673617786831326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,7168,0.007169778148333232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,2048,0.005591111050711737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,7168,0.02162844439347585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,6144,0.00777599960565567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,2048,0.005429333282841577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,1536,0.0064399999876817065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,5120,0.36367376645406085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,64,0.010674667027261523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,32,0.003168888803985384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,16384,0.07055022319157918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,16384,0.039112889104419284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,1024,0.0072266666425599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,5120,0.022153778208626643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,1024,0.005937777873542573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,256,0.011109333071443768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,2048,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,4096,0.03197777933544583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,3584,0.04894311229387919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,256,0.004019555532270008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,5120,0.011598221957683563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,3584,0.018964444597562153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,512,0.004649777793222004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,2560,0.035707556539111666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,512,0.0627431141005622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,512,0.003306666596068276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,4096,0.015825778245925903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,1536,0.026519111461109583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,7168,0.028170665105183918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,32,0.005780444376998478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,768,0.0037360000941488477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,6144,0.023853333459960088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,2560,0.009791110952695211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,128,0.018479110466109384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,512,0.023574221465322707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,4096,0.00792711145348019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,3584,0.1296995613310072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,512,0.005868444426192178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,12288,0.020004444652133517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,8192,0.026350221700138513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,3072,0.012295111185974546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,3072,0.019163555569118924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,4096,0.012315555579132505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,10240,0.05763022104899088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,768,0.037068443165885076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,16384,0.02437244521247016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,128,0.01428533262676663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,64,0.04245333207978142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,3584,0.018198221921920776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,256,0.008821333448092142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,12288,0.026339555780092876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,10240,0.012719111310111152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,3584,0.36126666598849827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,64,0.005576888720194499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,7168,0.02493600050608317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,5120,0.03197066651450263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,6144,0.03829066620932685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,2048,0.005982222242487802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,64,0.01646133263905843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,512,0.008058666355080074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,3584,0.007532444265153673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,3584,0.4782835642496745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,5120,0.013651555611027611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,12288,0.019235556324323017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,128,0.007859555383523306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,16384,0.543008910285102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,64,0.004473777694834603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,32,0.0025884444928831523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,1024,0.010303999814722273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,1536,0.014511111709806653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,6144,0.013730666703648038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,16384,0.18554400073157418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,1024,0.00868444475862715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,2560,0.009845332966910468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,16384,0.012225777738624148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,2560,0.00739644467830658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,1536,0.014505777094099255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,10240,0.053025778796937734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,2560,0.008915555146005418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,1536,0.00868800034125646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,512,0.009299555586444007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,768,0.00872444444232517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,7168,0.028192000256644353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,2048,0.006324444380071428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,64,0.025407110651334126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,128,0.002974222310715251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,768,0.003653333418899112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,5120,0.03141333328353034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,12288,0.0471937788857354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,6144,0.021144888467258875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,512,0.005987555616431766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,7168,0.016639999217457242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,2560,0.007177777588367462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,64,0.004652444273233414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,3072,0.015189333094490899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,512,0.0033617777129014335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,64,0.002773333340883255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,128,0.07168977790408664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,7168,0.06422844198015001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,512,0.005498666730191972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,10240,0.026366222235891554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,1536,0.0074924445814556545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,7168,0.01732711162832048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,256,0.015422221687104968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,5120,0.012700444294346703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,32,0.005008000052637524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,3072,0.5833759837680393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,128,0.004278222305907143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,10240,0.02516088883082072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,512,0.007820444802443186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,512,0.012977778083748288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,7168,0.03135733471976386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,128,0.002720000015364753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,6144,0.0188755558596717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,8192,0.028224888775083754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,3072,0.09534844424989487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,5120,0.07813244395785861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,3584,0.020255110330051847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,8192,0.011917333636018964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,3072,0.007369777394665613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,512,0.007515555454625024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,1024,0.004997333304749595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,12288,0.019934222102165222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,3072,0.0277164445983039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,1024,0.005149333427349727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,256,0.004384888956944148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,2048,0.01519377695189582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,768,0.008952889177534316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,8192,0.03154755632082621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,32,0.0025884444928831523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,2560,0.006395555618736479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,768,0.005723555468850666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,5120,0.04284266630808512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,12288,0.016862221890025668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,3072,0.008456889126035903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,16384,0.027462222509913977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,4096,0.012601777911186218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,1536,0.013445333474212222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,4096,0.008360889222886827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,12288,0.025263110796610515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,16384,0.09160533216264512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,2560,0.010281778044170803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,5120,16384,0.05053955647680494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,1536,0.01035288886891471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,12288,0.028252445989184912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,768,0.005277333160241445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,128,0.008567111359702216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,4096,0.07564355267418756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,5120,0.03542044427659776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,16384,0.5867999924553765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,3584,0.038482666015625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,2560,0.022011554903454248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,32,0.006311111152172089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,3072,0.00775999989774492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,32,0.0034542222403817703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,7168,0.06950755251778497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,2048,0.018651554981867473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,10240,0.03022222386466132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,1536,0.04412088791529337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,512,0.005407111098368962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,12288,0.008223111430803934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,256,0.002983111060327954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,12288,0.3104364342159695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,6144,0.01274400038851632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,65536,0.08491466442743938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,8192,0.17211822668711343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,2048,0.05991022454367744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,16384,0.030171556605233088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,16384,0.029829333225886028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,32,0.010803555448849997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,1024,0.005131555514203178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,1024,0.012828444441159567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,16384,0.02642400066057841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,3072,0.016207999653286405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,1024,0.008721777962313758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,3072,0.06435466474956937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,12288,0.10702755716111924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,8192,0.07802666558159722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,3584,0.11675822734832764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,256,0.0053022222386466126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,2048,0.019779556327395968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,7168,0.015107555521859063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,6144,0.01053422192732493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,2048,0.006796444455782573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,768,0.005620444400442972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,10240,0.020138666033744812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,4096,0.00907466643386417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,512,0.004433777597215441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,8192,0.01625777781009674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,64,0.00500266667869356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,1536,0.004276444514592488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,7168,0.012528888881206512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,1536,0.009887110855844285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,512,0.007893333004580604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,5120,0.015462223026487561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,1536,0.0640159977806939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,1536,0.02113155523935954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,128,0.02686311139000787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,65536,0.4826329019334581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,8192,0.013607111242082385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,768,0.004655111167165968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,768,0.014977777997652689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,1536,0.0046880000995265115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,768,0.010261333651012844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,256,0.007076444725195567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,512,0.006743999818960826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,1536,0.006656000183688269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,3584,0.03304266598489549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,3072,0.010648888846238455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,3072,0.00813066628244188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,32,0.0052382221652401825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,32,0.0053431110249625305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,65536,0.3783066537645128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,16384,0.026692445079485577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,256,0.0041911109454101985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,12288,10.191007826063368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,32,0.0032720000793536506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,1024,0.011873777541849347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,7168,0.018711111611790128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,7168,0.04606044292449951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,12288,0.20962488651275635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,16384,0.03266400098800659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,4096,0.01479200025399526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,64,0.0052382221652401825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,1024,0.05565777752134535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,10240,0.1257795492808024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,768,0.005501333210203383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,128,0.005214222189452913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,3584,0.007038222418891058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,1536,0.008997333546479544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,65536,0.1160400046242608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,2048,0.010954666468832227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,8192,0.012859555582205454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,16384,0.020968000094095867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,2048,0.008217777642938826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,10240,0.034188446071412824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,1536,0.005526222288608551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,64,0.00981866651111179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,32,0.00888000014755461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,3584,0.036293334431118436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,2560,0.03000533249643114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,768,0.006137777947717243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,5120,0.22958132955763075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,2560,0.007080888582600488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,4096,0.008360000120268928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,65536,0.13530577553643122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,128,0.0053137776752312975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,5120,0.011710222396585675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,4096,0.015984000431166757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,10240,0.14419289429982504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,3072,0.008476444416575963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,6144,0.03277333246337043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,128,0.004931555440028508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,256,0.0075039996041191956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,2048,0.011042666931947073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,7168,0.6911662419637045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,10240,0.007634667058785756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,1536,0.00756533361143536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,4096,0.01235466698805491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,1024,0.020770667327774894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,4096,0.02641600039270189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,2048,0.007876444194051955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,768,0.00573244442542394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,7168,0.011934222446547614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,1024,0.004843555390834808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,65536,0.4942604700724284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,256,0.033844444486829973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,6144,0.02405333353413476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,3072,0.03581422236230638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,16384,0.10345511304007636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,5120,0.013912889692518445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,2560,0.006174222048785951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,64,0.00721599989467197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,3584,0.007701333198282454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,8192,0.015596444408098856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,16384,0.023425777753194172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,2560,0.024502222736676533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,8192,0.030896888838873968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,1536,0.00849777791235182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,10240,0.049413333336512245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,4096,0.007567110988828871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,1536,0.00519822206762102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,4096,0.038513776328828596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,2560,0.006224888894293044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,6144,0.01796266602145301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,3584,0.03393955694304572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,3584,0.02946044339074029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,4096,0.016048888365427654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,12288,0.020145777199003432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,128,0.005079999979999331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,2560,0.012273777690198688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,32,0.0029440000653266907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,6144,0.015820443630218506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,8192,0.016137777103318107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,3072,0.02753866712252299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,8192,0.018710222509172227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,1024,0.003744889050722122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,2048,0.010241777532630498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,3072,0.016164445214801364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,4096,0.3160719871520996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,32,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,2560,0.4681190914577908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,7168,0.02875111169285244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,10240,0.017613333132531907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,1024,0.012162666353914471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,2048,0.015948444604873657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,7168,0.02407466702991062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,1536,0.014519999424616495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,1024,0.007599111232492659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,10240,0.02348444362481435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,64,0.002966222249799304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,1024,0.024708444873491924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,768,0.00777599960565567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,8192,0.011508444117175208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,512,0.03148533238304986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,64,0.010375111467308467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,128,0.003156444471743372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,1024,0.004197333421972063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,12288,0.015199111567603217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,128,0.004643555730581284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,5120,0.01108444482088089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,3072,0.05398222141795688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,12288,0.006588444527652528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,128,0.0028115556471877625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,3072,0.017218665944205392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,7168,0.24248088730706108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,256,0.0041288890772395665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,1536,0.008641777767075432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,10240,0.016436444388495553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,4096,0.5431529151068794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,6144,0.01722133325205909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,2048,0.4888888994852702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,3072,0.01611111064751943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,4096,0.07571555508507623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,2560,0.1324711110856798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,10240,0.015133332875039844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,16384,0.031055112679799397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,3072,0.04979733294910855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,6144,0.02504088812404209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,8192,0.027993778387705486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,6144,0.03947644432385763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,2560,0.03640177845954895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,3072,0.009387555221716562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,6144,0.2885173426734077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,768,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,8192,0.04996088809437222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,128,0.005205333232879639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,256,0.005017777697907554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,2560,0.23859998914930555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,6144,0.018208889497650992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,128,0.006488889041874144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,3584,0.013612444202105204
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,6144,0.03332711259524027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,10240,0.04634044567743937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,65536,0.263267543580797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,65536,0.23513510492112902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,12288,0.026583999395370483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,768,0.0038764443662431505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,1536,0.009333333207501305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,512,0.026546667019526165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,16384,0.030176000462638006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,5120,0.07391733593410917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,16384,0.0314684444003635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,64,0.002568888788421949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,512,0.007213333414660559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,7168,1.1727608574761283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,8192,1.601576911078559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,4096,0.014278221461508008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,8192,0.02907466557290819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,768,0.010391111175219217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,6144,0.00811288919713762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,65536,0.20395733250512016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,768,0.01126933346192042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,6144,0.02850666642189026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,256,0.0033884443756606844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,2560,0.012615111139085559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,3072,0.14454221725463867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,65536,0.04035377833578322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,768,0.02331111166212294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,5120,0.0069066666894488865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,2048,0.012673777838548025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,7168,0.01016977760526869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,6144,0.006280888699822956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,768,0.006814222368929122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,128,0.002817777709828483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,32,0.002562666725781229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,32,0.002489777695801523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,7168,0.013416888813177744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,1024,0.00959644466638565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,32,0.005306666509972678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,6144,0.013375110924243927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,3072,0.007885333564546373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,4096,0.014612444572978549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,128,0.021065778202480737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,2560,0.005956444475385878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,512,0.005899555567238066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,16384,0.05650133556789822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,3072,0.015047111445003085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,2560,0.0064186664919058485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,7168,0.010951999988820819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,1024,0.0064311110311084324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,3584,0.008045333127180735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,768,0.006154666758245892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,6144,0.023035556077957153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,64,0.0029884444342719186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,2560,0.022064889470736187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,768,0.003921777837806278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,65536,4096,0.07692088683446248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,1536,0.01071466671095954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,512,0.005156444592608346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,4096,0.016773333152135212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,3072,0.06108533011542427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,8192,0.027494221925735474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,16384,0.09118222528033787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,7168,0.020780444145202637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,512,0.0033528889632887314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,2560,0.012120000190205045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,512,0.01945244438118405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,16384,0.011693333586057028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,6144,0.32751644982231987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,12288,0.015079999963442484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,3072,0.014187556174066333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,768,0.008392000363932716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,64,0.005260444349712796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,7168,0.016511110795868766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,512,0.0063288890653186375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,6144,0.01366133325629764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,32,0.0038817777401871155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,65536,0.10671644740634495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,6144,0.13105866644117567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,65536,1.3380230797661676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,1024,0.006760889043410619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,6144,0.02514311174551646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,2048,0.03357422351837158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,768,0.01277599980433782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,7168,0.02795377704832289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,3072,0.03712000118361579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,256,0.005727111051479976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,65536,0.07845422294404772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,10240,0.007568000091446771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,4096,5120,0.01698844465944502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,65536,0.9150595135158963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,5120,0.02238133384121789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,12288,0.014470222923490735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,768,0.012786666552225748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,512,0.08237422174877591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,7168,0.01628266606065962
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,3072,0.011543999943468304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,8192,6.899179670545791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,2560,0.058800889386071094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,16384,0.010251555177900527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,3584,0.007890666524569193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,1536,0.00481244424978892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,2048,0.010843555960390302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,32,0.0057706667317284485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,8192,0.04698222213321262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,64,0.008600888980759515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,64,0.006793777975771163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,5120,0.1838444471359253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,6144,0.010991111397743225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,128,0.0027679999669392905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,2048,0.005662222289376789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,12288,0.13106577926211888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,768,0.024148444334665935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,768,0.005480888817045424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,2560,0.005986666513813867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,3584,0.012877333495352002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,8192,0.011895111037625207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,6144,0.13567199971940783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,2048,0.017108443710539076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,128,0.07122221920225355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,16384,0.14615021811591253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,1536,0.05565155877007378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,256,0.0053102220926019884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,8192,0.016406221522225272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,6144,0.011117333339320289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,64,0.0028942223224374983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,10240,0.006583111153708563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,2048,0.008060444560315874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,7168,0.050887998607423574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,512,0.0058266665372583605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,3584,0.01366311146153344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,3072,0.05799377626842923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,65536,1.0147519641452365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,3072,0.008740444150235917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,2048,0.009471110999584198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,1024,0.009413333402739631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,4096,0.015214223000738354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,10240,0.05310044354862637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,8192,0.07952977551354302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,32,0.002733333243264092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,12288,0.04461155666245354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,65536,0.22263201077779135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,32,0.0027262222849660446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,2048,0.02532800038655599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,7168,0.013351110948456658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,32,0.002720000015364753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,65536,0.07090044683880277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,768,0.0046746668716271715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,512,0.009263111485375298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,7168,0.018868444694413077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,768,0.010302222437328762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,3072,0.014030221435758801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,3072,0.009425777528021071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,512,0.004447999927732679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,4096,0.006110222389300664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,12288,0.139265775680542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,128,0.00564177789621883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,128,0.0029893333299292457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,256,0.005224888937340842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,7168,6144,0.02490577763981289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,256,0.017888888716697693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,5120,0.014323555760913424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,5120,0.008960000342792934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,7168,0.13099111451043025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,12288,0.050311999188529126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,768,0.00793333351612091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,2560,0.006483555667930179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,12288,0.0151573336786694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,512,0.004443555656406614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,5120,0.17131643825107148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,2048,0.010464888479974536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,256,0.08147733079062568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,32,0.08880711264080471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,256,0.006768888897365994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,2560,0.006480000085300869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,7168,0.01294844432009591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,3072,0.010865777730941772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,3072,0.006520000182920032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,128,0.006229333165619109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,4096,0.18301867114173043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,32,0.005769777629110549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,4096,0.010474666953086853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,512,0.0037288889288902283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,65536,0.06948088937335543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,1536,0.00944088896115621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,256,0.011783999701340994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,2048,0.011108444796668159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,4096,0.01109955542617374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,5120,0.02957688768704732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,12288,0.005912000106440649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,12288,0.018904000520706177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,512,0.004504888835880491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,256,0.006548444430033366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,2048,0.010798222488827176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,1024,0.005345777918895085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,16384,0.022425777382320825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,4096,0.029096888171301946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,8192,0.008390222158696916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,8192,0.020503999458418954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,3584,0.015281778242852954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,256,0.004394666602214177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,3584,0.006501333167155583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,16384,0.16447822252909342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,10240,0.010287111004193624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,3584,0.05458133419354757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,10240,0.18101333247290718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,2048,0.013215111361609565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,128,0.014331556028789945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,768,0.006572444405820634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,6144,0.010114666488435533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,65536,0.14966400464375815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,768,0.0053182223604785065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,3584,0.006768888897365994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,12288,0.16556089454227024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,16384,0.5100613170199925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,2048,0.005559111220969095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,64,0.0038053331275780997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,1024,0.006917333437336816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,6144,0.009640889035330879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,12288,0.05533155467775133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,64,0.025880000657505457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,1024,0.004850666556093427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,16384,0.10293244653277928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,512,0.008034666379292807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,256,0.005373333477311664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,3072,0.03344622254371643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,3584,0.008443555898136562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,5120,0.01183555523554484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,2560,0.007238222493065729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,3072,0.007888000044557784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,3584,0.017137777474191453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,512,0.00959022260374493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,768,0.003444444388151169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,128,0.0037546666959921518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,2560,0.009220444493823582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,4096,0.019146665930747986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,12288,0.029311110575993855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,1024,0.009315555294354757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,1536,0.02163733376397027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,2048,0.013552889227867126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,1024,0.0057857777509424425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,5120,0.03382222188843621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,256,0.003129777808984121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,5120,0.013986666997273764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,2048,0.06603555546866523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,32,8192,0.011039111349317761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,768,0.004341333276695675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,3584,0.09470578034718831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,65536,0.1526906622780694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,5120,0.01147911118136512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,1024,0.00535022219022115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,16384,0.02014844450685713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,2560,0.0069857777820693115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,6144,0.010668444136778513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,2560,0.007888889147175683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,512,0.003579555700222651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,65536,0.0750711096657647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,64,0.0025884444928831523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,64,0.006680000159475539
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,512,0.009372444616423713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,768,0.3638906743791368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,10240,0.020124443703227572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,4096,0.03377866744995117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,7168,0.0690880020459493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,6144,0.013054221868515015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,128,0.004000888930426704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,768,0.014892444014549255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,128,0.037718223200904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,32,0.0037377778854635027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,12288,0.3734009000990126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,10240,0.022422222627533808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,7168,0.016695110334290397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,1024,0.013779555757840475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,1024,0.9425546858045789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,1536,0.007706666986147563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,6144,0.01181244436237547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,768,0.007345777418878343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,256,0.010108444425794814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,512,0.004371555729044808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,768,0.004484444442722532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,512,0.01660977800687154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,2048,0.042555554045571216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,3072,0.00790666706032223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,1536,0.007842666572994655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,7168,0.15196800231933594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,8192,0.009152889251708984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,12288,0.04683644572893778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,65536,0.18511822488572863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,3584,0.01423822177780999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,10240,0.01697422232892778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,512,0.21720533900790742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,2048,0.025172443853484258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,5120,0.05309333403905233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,768,0.00507377791735861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,1024,0.007269333634111617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,5120,0.021328889661365088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,1024,0.4346755610571967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,65536,0.1240524450937907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,4096,0.006578666468461354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,8192,0.027853333287768896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,1536,0.005509333478079901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,3584,0.02370311154259576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,128,0.005008000052637524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,768,0.006069333189063602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,10240,0.013979555832015144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,12288,0.04061955544683669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,5120,0.010223999619483948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,8192,0.02141155633661482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,6144,0.012674666941165924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,16384,0.023073777556419373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,8192,0.391762653986613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,512,0.005675555517276128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,2048,0.005201777650250329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,1536,0.0047102222839991255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,6144,0.015292444162898593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,12288,0.05742310815387302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,4096,0.035094221433003746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,256,0.005503111001518037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,64,0.005903111149867375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,1536,0.012828444441159567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,256,0.0032417778339650896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2560,3072,0.010088000032636855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,6144,0.014666666587193808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,512,0.0033404444240861465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,1536,0.14568177858988443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,4096,0.015484443969196744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,65536,0.062209778361850314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,10240,0.019888889458444383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,32,0.005227555417352253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,3584,0.01797155539194743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,6144,0.22945332527160645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,1024,0.15093333191341823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,3072,0.007630222373538547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,10240,0.006416888700591193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,3584,0.050056000550587974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,2560,0.035438223017586604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,16384,0.007812444534566667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,6144,0.012739555703269111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,4096,0.38158133294847274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,256,0.004912000149488449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,1024,0.0068400001360310455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,2048,0.010512000156773461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,2048,0.005294222384691238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,65536,0.012316444681750404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,32,0.003144888828198115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,7168,0.12057244777679443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,12288,0.018087110585636564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,3584,0.01332711097266939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,32,0.002700444517864121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,512,0.009661333428488838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,1536,0.008737777670224508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,256,0.007126222054163615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,768,0.09067822164959377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,2560,0.010601777997281818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,2560,0.007962666451931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,128,0.0032622222271230486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,2560,0.009493333597977957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,65536,0.8860746489630805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,32,0.0028151110228565005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,32,0.0038204445607132385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,3072,0.01292799992693795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,32,0.009648888475365108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,256,0.003510222252872255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,5120,0.008709333009190029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,2048,0.009141333401203156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,2048,0.007363555332024892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,256,0.0032888888898822996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,1024,0.005792888916201062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,5120,0.021715554926130507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,4096,0.012158222496509552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,16384,0.01035733355416192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,768,0.003956444561481476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,16384,0.21696000629001191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,768,0.005467555589146084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,16384,0.023046221998002794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,4096,0.008459555606047312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,16384,0.020404444800482858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,768,0.006138666636413998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,512,0.010203555226325989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,8192,0.016766221986876596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,16384,0.7284364700317383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,8192,0.01461955573823717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,512,0.013022222452693515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,4096,0.010482666393121084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,1024,0.008429333567619324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,128,0.009293333523803288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,3584,1.4621689054701064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,3072,0.008954666554927826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,1024,0.01426488823360867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,16384,0.014071111877759298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,1536,0.05065066615740458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,5120,0.05003822180959913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,256,0.0034808889031410217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,2048,0.006346666564544042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,1024,0.004000888930426704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,8192,0.023257777094841003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,2048,0.008753777378135258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,512,0.003963555726740095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,768,1536,0.00666133314371109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,768,0.007678222325113084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,64,0.002740444408522712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,10240,0.016083555089102853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,8192,0.016736000776290894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,8192,0.03195377853181627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,512,0.0033164444482988785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,1024,0.004970666848950916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,6144,0.02351288828584883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,1536,0.011888888974984487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,1536,0.24506399366590711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,512,0.019113777412308585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,3072,0.012847111456924014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,2048,0.012142221960756512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,128,0.004357333398527569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,4096,0.006216889040337668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,4096,0.006228444476922353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,256,0.00498933345079422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,4096,0.007870222131411234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,2048,0.021735111872355144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,128,0.0041271112859249115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,512,0.003678222083383136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,3072,0.006535999890830781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,12288,0.01588622232278188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,12288,0.03294577863481309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,3072,0.0064488889442549805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,32,0.004936000125275718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,2048,0.008723555339707268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,128,0.006256888724035687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,12288,0.08909422159194946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,16384,0.1433360046810574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,64,0.006011555592219035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,64,0.0031591111587153543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,10240,0.02494844463136461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,128,0.005121777868933148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,6144,0.011880888707107969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,3584,0.029055999384986028
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,4096,0.02306488818592495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,12288,0.09238577551311916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,128,0.005009777843952179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,12288,0.026348445150587294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,65536,0.1354328923755222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,2048,0.012791111237472959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,128,0.004899555610285865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,512,0.00517155561182234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,7168,0.014467555615637036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,64,0.013624000052611033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,768,0.005672889037264719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,2048,0.19467200173272026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,512,0.013691555294725629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,512,0.006872889068391588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,10240,0.02606755495071411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,3072,0.012331555287043253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,2560,0.013941332697868347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,10240,0.04062666826777988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,2560,0.015461333923869662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,12288,0.022451554735501606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,4096,0.008811555802822113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,4096,0.01092355532778634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,6144,0.016453334026866488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,7168,0.012669333153300814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,12288,0.09362044599321152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,12288,0.47891998291015625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,2560,0.01107022249036365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,5120,0.020640888147883944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,3584,0.012681778106424542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,128,0.004468444320890638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,12288,0.030177778667873804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,768,0.14844444062974718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,10240,0.7958417998419868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,7168,0.1849306689368354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,7168,0.029254221253924902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,2048,0.008657777474986183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,512,0.005889777921968036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,65536,0.10669866535398696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,3072,0.008441777692900764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,256,0.003133333391613431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,3584,0.009948444035318162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,10240,0.028371555937661067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,2048,0.0076044441925154785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,512,0.004208000169859992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,6144,0.03577866819169786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,3072,0.007881777981917063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,65536,0.3836497730678982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,1024,0.03012711140844557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,1024,0.006713777780532837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,65536,5.935258653428819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,65536,0.1496773295932346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,1536,0.019457777341206867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,8192,0.018889778190188937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,5120,0.0170168893204795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,256,0.017475555340449016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,2048,0.04380089044570923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,64,0.0027599999060233435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,768,0.0058044443527857465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,65536,0.14920800262027317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,128,0.004407999830113517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,7168,0.02085955606566535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,10240,0.030001779397328694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,1536,0.005640000104904175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,1024,0.005543111099137201
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,128,0.006640888750553131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,1024,0.005110222018427319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,32,0.002874666617976295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,1024,0.02089333368672265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,768,0.008199111455016667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,16384,0.03866844375928243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,2048,0.07191289133495755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,65536,0.16018133693271214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,768,0.008160889148712158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,16384,0.03245066603024801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1536,12288,0.017496888836224873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,6144,0.01238222254647149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,12288,0.994201766120063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,4096,0.01372977760103014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,8192,0.03328088919321696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,768,0.005664888769388199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,16384,0.07235822412702772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,32,0.0071928890215026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,16384,0.008640000389681922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,2560,0.012568888564904531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,5120,0.36156267589992946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,10240,0.006401777681377199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,2048,0.008639111287064022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,4096,0.08854489194022284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,8192,0.014538667268223233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,2048,0.005604444278611078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,512,0.008079999850855933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,512,0.03133066826396518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,6144,0.1118968857659234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,3584,0.011333333121405708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,8192,0.018918222851223417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,6144,0.02260977857642704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,16384,13.601497226291231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,8192,0.04718755682309469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,6144,0.031255112753974065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,128,0.007637333538797166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,64,0.005469333380460739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,2048,0.014277334014574686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,128,0.009474666582213508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,2560,0.008383110993438298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,8192,32,0.0029288888391521242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,8192,0.07843910985522799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,3584,0.01331733332739936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,1536,0.016122667325867545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,64,0.0058542220956749385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,2048,0.06829866435792711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,256,0.004233777936961916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,3072,0.012450666891203986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,512,0.0155395550860299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,32,0.005383111122581694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,7168,0.8729235331217448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,2048,0.007088888850477006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,512,0.005935110979610019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,32,0.00847466703918245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,3584,0.010249777800507015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,3072,0.006615110983451207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,4096,0.009665778113736046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,6144,0.20267021656036377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,1536,0.009979555176364051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,1536,0.008893333375453949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,128,0.003096888876623578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,32,0.0026844443960322272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,128,0.008008889026112026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,128,0.0038764443662431505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,512,0.007121777368916406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,64,0.003026666740576426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,4096,0.021927111678653296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,3584,0.01237777786122428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,3584,0.016904888881577384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,10240,0.036659555302725896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,256,0.004381333374314838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,64,0.009069333473841349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,1024,0.11673422654469807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,768,0.09911733203464085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,32,0.0052133335007561584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,8192,0.02011999984582265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,768,0.005477333234416113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,128,0.005331555588377847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,768,0.004252444538805219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,2048,0.03656355539957682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,128,0.0031306667046414483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,7168,0.011387555963463254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,1024,0.00416355538699362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,65536,0.14555554919772679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,8192,0.02932000160217285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,512,0.00721866637468338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,768,0.019419555862744648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,2560,0.0513626668188307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,64,0.003031999907559819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,10240,0.021528888079855178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,768,0.005856888989607493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,4096,0.012947555217478009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,12288,0.02432800001568264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,256,0.006631111105283101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,3584,0.007408888803588019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,6144,0.012801777985360889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,2560,0.007254222200976477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,65536,0.01460355520248413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,10240,0.04051911168628269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,256,0.00756444450881746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,5120,0.019647111495335896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,32,0.0053031109273433685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,768,0.00794044468137953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,64,0.0025742221623659134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,1536,0.009729777773221334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,128,0.00499555551343494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,768,0.00507377791735861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,3072,0.01147822207874722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,768,0.0036559998989105225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,65536,0.07580000162124634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,6144,0.11525777975718181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,16384,0.02127200033929613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,1536,0.005507555686765247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,2048,0.009292444421185387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,32,0.005580444302823808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,65536,2.397030300564236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,16384,0.026914666096369427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,16384,0.028862221373452082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,10240,0.015674667225943673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,64,0.005475555443101459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,3072,0.007430222299363878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,3584,0.008384000096056197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,1536,0.013220444321632385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,2048,0.013088000317414602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,2560,0.2583306630452474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,3584,0.007893333004580604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,10240,2.0819945865207248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,512,0.004639111045334074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,16384,0.07503022087944879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,10240,0.2839164468977186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,256,0.004412444515360726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,16384,0.03711288836267259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,3072,0.008064000142945183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,256,0.0053591111467944244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,65536,0.2314213381873237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,5120,0.052321778403388135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,64,0.005980444451173146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,5120,0.009454222189055549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,16384,0.02630311085118188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,1536,0.007872888611422645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,7168,0.010996444357766045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,256,0.003087999920050303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,3584,0.008408889174461365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,10240,0.018194667167133756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,3072,0.01886044442653656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,16384,0.05147644546296862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,16384,4.1633555094401045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,2048,0.11898844771915013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,16384,0.025901334153281316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,5120,0.027549333042568628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,3584,0.0722000002861023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,64,0.00721244431204266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,768,0.01071999967098236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,3584,0.008610666626029545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,64,0.03685777717166477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,1536,0.005466666486528184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,1024,0.011680888632933298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,1536,0.010278222461541494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,512,64,0.005134222408135732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,7168,0.026908444033728704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,7168,0.011016888750924004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,768,0.0058124446206622655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,512,0.0031146665828095544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,256,0.002830222249031067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,4096,0.0076657773719893555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,8192,0.015847999188635085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,10240,0.033589333295822144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,32,0.008347555167145198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,128,0.0038284444146686127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,65536,0.9041973749796549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,32,0.003101333354910215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,1536,0.012744888663291931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,1024,0.006355555521117316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,10240,0.10068266921573216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,768,0.006107555495368109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,512,0.012181333369678922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,12288,0.41027379035949707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,8192,0.3939208984375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,12288,0.0564017759429084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,5120,0.046463999483320445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,2048,0.06334488921695285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,2560,0.019381332728597853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,2560,0.00775911079512702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,128,0.039797332551744245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,6144,0.035353776481416486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,5120,0.042618665430280894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,7168,0.028143111202451918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,1024,0.027021333575248718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,1536,0.015192000402344598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,16384,0.05511377917395698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,10240,0.018806222412321303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,3072,0.006721777634488211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,6144,0.025161777933438618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,16384,0.3929608927832709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,8192,0.014377777775128683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,128,0.005144000053405762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,256,0.005562666803598404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,32,0.0026737778551048706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,3584,0.013433777623706393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,6144,0.041907555527157254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,2560,0.026476444469557867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,4096,0.007990222010347579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,16384,0.02402755618095398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,4096,0.054450667566723294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,6144,0.05374933282534281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,5120,0.052779555320739746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,2048,0.005395555661784277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,512,0.005510222166776657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,768,0.003908444609906939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3584,1024,0.006117333140638139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,6144,0.006687110910813014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,32,0.0028817777832349143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,5120,0.014243554737832812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,3584,0.008290666672918532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,7168,0.03055644366476271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,1024,0.012147555748621622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,65536,0.01256444470749961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,768,0.006180444525347815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,7168,0.06428444385528564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,3072,0.00959822204377916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,1024,0.011853333148691388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,3072,0.012335110869672565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,4096,0.03378666771782769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,512,0.003724444243643019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,4096,0.01889155473974016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,2560,0.010800888968838586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,6144,0.01758577757411533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,4096,0.13010666105482313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,256,1024,0.005903999838564131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,2048,0.008576889004972246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,65536,0.2829102145300971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,65536,0.7170009083218045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,65536,0.2924942175547282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,128,0.0064399999876817065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,128,0.008022222254011366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,6144,0.007361777954631382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,1536,0.012665777570671506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,8192,0.015761777758598328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,64,0.002979555477698644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,7168,0.0107351111041175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,512,0.05009777678383721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,16384,0.007693332930405934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,5120,0.8347324265374078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,65536,4.804857889811198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,5120,0.04373155699835884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,2048,0.011140444212489657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,2048,0.009898666706350114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,512,0.0075395554304122925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,128,0.005145777844720417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,10240,0.0228426671690411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,4096,0.010430222584141625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,32,0.002620444529586368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,1024,0.06354488929112752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,2048,0.006877333339717653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,7168,0.036256889502207436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,512,0.004457777573002709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,512,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,2048,0.02346044447686937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,8192,0.018608000543382432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,10240,0.017444445027245414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,12288,0.01754844519827101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,1024,0.008721777962313758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,2560,0.007881777981917063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,64,0.002951111023624738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,1536,0.009852444132169088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,3584,0.009879999690585667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,2048,0.005981333139869902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,3072,0.012103111379676394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,6144,0.10964800251854791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,128,0.009466666314336989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,6144,0.5512888696458604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,768,0.007831999825106727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,64,0.0025902222841978073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,10240,0.01534222231970893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,1536,0.0063182223174307085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,7168,0.02516355613867442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,10240,0.0159262220064799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,256,0.006480000085300869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,3584,0.12649333477020264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,1024,0.005460444423887465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,3072,0.013941332697868347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,1024,0.018388445178667705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3072,10240,0.019824888971116807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,512,0.0032995556377702286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,7168,0.02782755427890354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,7168,0.017667555146747164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,4096,0.006405333264006509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,2560,0.007820444802443186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,5120,0.007264888948864407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,128,0.0027155555370781156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,8192,0.29905690087212455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,1536,0.005952000204059813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,256,0.02164622147878011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,7168,0.01572177807490031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,8192,3.320427576700846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,8192,0.2939697901407878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,1024,0.008749333520730337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,10240,0.09213422404395209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,3072,0.009205333060688442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,256,0.0038684445122877755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,16384,16384,0.12532088491651747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,6144,0.03792000148031447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,5120,0.014097778333557976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,65536,0.012680889003806643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,32,0.002603555512097147
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,4096,0.012490666574902005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,768,0.004874666531880696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,3584,0.018241778016090393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,7168,0.006609777609507243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,3584,0.009783999787436591
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,3072,0.0206666671567493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,12288,12.416568332248263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,3584,0.025158221522967022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,512,0.028079999817742243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,8192,0.032223112053341336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,2048,0.007745777567227681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,256,0.005782222168313132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,12288,0.037520888778898455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,10240,0.013895110951529609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,64,0.006121777825885349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,12288,0.036181333992216326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,1024,0.012772444221708508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,768,0.049957331683900624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,6144,0.05438666542371114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,4096,0.009450666606426239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,10240,0.3283209005991618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,10240,0.020962667134073045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,1536,0.0069351109365622205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,12288,0.020095111595259774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,768,0.06236444579230415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,2560,0.022164444128672283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,6144,0.013593778014183044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,3584,0.018138666947682697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,1536,0.007367110914654202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,512,0.0034133332471052804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,5120,0.04174133472972446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,2560,0.022387555903858606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,7168,0.09241777658462524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,8192,0.022150221798155043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,5120,0.01629955569903056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,12288,0.044870220952563815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,65536,0.8121004634433322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,65536,0.01346666696998808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,6144,0.47053247027926975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,5120,0.010777778095669217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,128,0.007197333706749811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,3072,0.020541333489947848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,3072,0.008353778057628209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,512,0.012638222012254926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,32,0.0058142224119769196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,2048,0.00997155573632982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,65536,0.20295911365085176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,3072,0.05781244569354587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,64,0.002806222273243798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,64,0.006274666637182236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,2048,0.006751110984219446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,8192,0.012916444076432122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,1024,0.008425777984990014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,32,0.002591111179855135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,4096,0.006160000132189856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,1024,0.003976888954639435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,3584,0.011955555942323474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,8192,0.04091644287109375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,2048,0.005509333478079901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,1024,0.005707555760939916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,65536,0.06617422236336602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,10240,0.014339556296666464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,1024,0.00490488898422983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,2560,0.007576888634098901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,6144,0.006413333531883027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,768,2048,0.005496888938877318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,7168,0.0532533327738444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,3072,0.06364533636305067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,10240,8192,0.027858667903476294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,2048,0.00654666663871871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,1024,0.008054222497675154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,3072,0.01795111099878947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,768,0.03349511159790887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,2048,0.013432000246312885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,2048,0.012916444076432122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,768,0.02843911117977566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,4096,0.00905777762333552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,3584,0.00909511082702213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,16384,0.055495109823015004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,5120,0.08877422412236531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,128,0.004573333180612988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,7168,0.042471110820770264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,1536,0.005538666827811136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,6144,0.04092533389727274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,32,0.008007999923494127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,65536,0.05941422118080986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,256,0.00295733329322603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,32,0.0036924444139003754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,768,0.012313777373896705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,16384,0.01833599971400367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,768,0.006107555495368109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,12288,0.01776000029510922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,8192,1.6615244547526042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,32,0.003699555579158995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,8192,0.015440889530711703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,12288,0.01663555536005232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,6144,0.15950132740868464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,256,0.014837332897716098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,512,0.007061333292060428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,2560,0.011391110718250275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,10240,0.03167377909024557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,32,0.0031822222388452957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,64,0.0028195555011431375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,3072,0.011526222030321756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,2560,0.0074782222509384155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,4096,0.07104177607430352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,5120,0.4488159815470378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,6144,0.2730026774936252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,7168,0.03065155612097846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,2048,0.011516444385051727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,4096,0.01885511146651374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,128,0.20399822129143608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,2560,0.015399111641777886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,65536,0.4156408839755588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,2048,0.08571022086673313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,12288,0.036328001154793635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,4096,0.007084444165229797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,768,0.006604444649484422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,3584,0.008731555607583787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,2560,0.01240533341964086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,2048,0.03797600004408095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,2048,0.10889689127604167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,512,0.007726222276687622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,512,0.008340444829728868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,1536,0.006194666855865055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,768,0.02216888964176178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,7168,0.02955377764172024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,512,0.00443466669983334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,512,0.005436444448100195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,1024,0.0046986668474144405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,1536,0.019759999381171334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,128,0.038422223594453596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,3072,1.115093337164985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,12288,0.03303111261791653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,65536,0.04333422250217862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,6144,0.009472889204819998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,32,0.003866666720973121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,4096,0.0284862220287323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,128,0.002936888900068071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,7168,0.010767111347781287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,256,0.004680888934267892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,32,0.005209777918126848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,256,0.005435555759403441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,5120,0.018766222728623282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,768,0.011305777562989129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,512,0.00332266671790017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,3072,0.013594666288958656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,1024,0.014887111054526435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,4096,0.011167999770906238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,12288,0.12451910972595215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,10240,0.032461331950293645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,3584,0.021408889028761122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,6144,0.5648960007561578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,2048,0.009679999616410997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,6144,0.11411110560099284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,2048,0.00960977789428499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,2048,0.007114667031500075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,7168,0.06367199950748019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,4096,0.03596977723969354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,2560,0.00795733349190818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,12288,0.234827544954088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,5120,0.011740444435013665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,768,0.01370666672786077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,768,7168,0.012057777908113269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,2048,0.02110044492615594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,3072,0.022502221994929846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,768,0.00360355567600992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,8192,0.011576000187132092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,4096,0.00851466672288047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,8192,0.0387368897596995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,32,0.055471999777687915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,64,0.005603555589914322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,12288,0.06472444534301758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,3584,0.009854222337404886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,2560,0.007010666860474481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,65536,0.38971911536322695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,12288,0.01455200049612257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,2560,0.010263111028406356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,7168,0.31305334303114146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,2048,0.09302133321762085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,1536,0.0788373351097107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,512,0.006949333681000604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,3584,0.1619982189602322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,128,0.00997688869635264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,128,0.003927111211750242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,512,0.040927112102508545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,1024,0.009447111023796929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,3584,0.01646844380431705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,768,0.005649777750174205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,1536,0.005481777919663324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,1536,0.009098666409651438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,65536,0.04506488972240024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3584,3072,0.010303110712104374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,3584,0.007752888732486301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,3072,0.02516266703605652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,256,0.003540444291300244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,128,0.0028071111689011254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2560,7168,0.01344622257683012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,256,0.0041315555572509766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,2560,0.02440622283352746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,128,0.0029377777957253983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,12288,0.01605333387851715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,6144,0.39958845244513613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,65536,0.18419110774993896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,4096,0.007320889168315464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,16384,0.038615110847685076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,32,0.002760888801680671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,8192,0.07606666617923312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,64,0.004473777694834603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,4096,0.011996444728639392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,128,0.002768888862596618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,6144,0.01035377797153261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,3584,0.008187555604510838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,7168,0.00869422240389718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,8192,0.15667111343807644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,32,0.0028826666788922418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,64,0.002687999978661537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,16384,0.030206223328908283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,65536,0.5422408845689561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,768,0.00941066692272822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,256,0.006986666884687211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,512,0.008362666600280339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,10240,0.034521778424580894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,512,0.005398222141795688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,2560,0.01687111126052009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,6144,0.06874489121966891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,16384,0.08106222417619494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,16384,0.11863644917805989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,4096,0.0064026667839950984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,256,0.004980444494220945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,3584,0.009527111219035255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,768,0.009650666680600908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,1536,0.07851022481918335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,768,0.0531217787000868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,1024,0.03755022088686625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,3072,0.011764444410800934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,8192,0.008660444782839881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,3072,0.006729777902364731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,3072,0.02380977736579047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,12288,7168,2.077967961629232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,10240,0.03918933206134372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,3584,0.014544889330863953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,1536,0.01534577707449595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2048,2560,0.008665777742862701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,128,0.0037902221083641052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,7168,0.024591111474566992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,4096,0.0069679998689227635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,12288,0.01943644384543101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,2560,0.38330401314629453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,4096,0.015599111715952555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,2560,0.023766222927305434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,3584,0.01707466608948178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,1536,0.03693155447642008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,10240,0.007517333659860823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,16384,0.08107999960581462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,512,0.01294311136007309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,5120,0.011186666786670685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,1536,0.013456888496875763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,5120,0.021039111746682063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,2560,0.373166216744317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,256,0.01758044461409251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,1536,0.005576000031497743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,1536,0.011861333416567909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,1024,0.009015999734401703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,768,0.005704000178310607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,12288,0.01900888813866509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,65536,0.07107200225194295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,12288,512,0.0446648895740509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,2048,0.009205333060688442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,5120,0.01829777823554145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,6144,0.04445866743723551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,512,0.005346666607591841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,768,0.006249777972698212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,2560,0.0074826669361856245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,10240,0.08978222476111518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,256,0.012250666817029318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,256,0.004872888740566042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,2560,0.006124444305896759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,1024,0.007747555772463481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,768,0.041834665669335254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,5120,0.008815111385451423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,16384,0.5041768815782335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,512,0.009654222263230218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,4096,0.06491377618577746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,64,0.005883555445406172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,128,0.007538666327794393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,1536,0.006583111153708563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,32,0.0034479999707804788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,12288,0.017157332764731515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,256,0.0033528889632887314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,2048,0.007524444825119442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,2560,0.007165333463086023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,3584,0.024383111132515803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,2560,0.07753155628840129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,256,0.0058044443527857465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,8192,0.25005777676900226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,6144,0.1095022228029039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,1024,0.004243555582231945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,8192,0.03860177927547031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,1536,0.006769777586062749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,768,0.0037840000457233856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,4096,0.012186666329701742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,6144,0.04016800059212579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,5120,0.033590220742755465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,5120,0.021169778373506334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,32,0.00254488881263468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,5120,0.13976977931128606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,1024,0.006292444550328785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,10240,0.026383999321195815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,768,0.0036640001667870414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,1024,0.006680888848172293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,512,0.012490666574902005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,65536,0.13370845052931044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,2048,0.012981333666377597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,768,0.0074275558193524676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,65536,0.21102310551537407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,6144,0.010975110861990186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,2560,0.006060444646411472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,65536,0.017582222819328308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,4096,0.005882666756709416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,1536,0.01000711073478063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,1536,0.006218666831652324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,1024,0.005017777697907554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,1536,0.010777778095669217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,256,0.003169777699642711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,4096,0.04866666595141093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,4096,0.18879289097256133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,16384,0.04787466592258877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,6144,0.018890667292806838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,64,65536,0.8601822323269315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,6144,0.02187999917401208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,12288,0.01777777738041348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,7168,0.05497511227925619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,4096,0.016754666964213055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3072,128,0.005554666535721884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,12288,0.005897777775923411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,1024,0.03254488772816128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,3072,0.010773333410422007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,7168,0.026711111267407734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,256,0.004391111019584867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,7168,0.010451555252075195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,8192,0.2692017820146349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,65536,0.14900888337029353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,1536,0.005049777941571342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,128,0.0029066666546795103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,5120,0.008255999949243333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,65536,0.1615084409713745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,4096,0.017631111873520743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,5120,0.007087111473083496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,3584,0.018040888839297824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,16384,0.02605777813328637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,128,0.005018666800525454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,7168,0.03509599963823954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,16384,0.008851555486520132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,65536,0.6468862427605523
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,128,0.004187555362780889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,3072,0.010934222075674268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,16384,0.023631110787391663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,65536,0.28605956501430935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,32,0.003142222141226133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,16384,0.021742223037613764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,8192,0.04930222365591261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,16384,0.019205333458052743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,12288,0.007970666719807519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,2560,0.007792000141408708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,32,0.07612444294823541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,6144,0.03430044319894578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,16384,0.02568088968594869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,12288,0.020310221446885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,512,0.004872888740566042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,3584,0.007683555285135905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,2560,0.006005333529578314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,512,0.004249777644872665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,2048,0.011271110839313932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,6144,0.013823111024167804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,3072,0.00998488896422916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,4096,0.021849777963426378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,3072,0.006519111080302133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,8192,12288,0.03363022208213806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,1536,0.013711999687883588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,3584,0.011076444553004371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,3072,0.03289955523278978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,2048,0.013749333719412485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,12288,0.09234222438600327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,2048,0.005537777725193236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,768,0.004597333156400257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,512,0.005103111267089844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,768,0.0086986662613021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,3072,0.012647111382749347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,256,0.003026666740576426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,512,0.005079111291302575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,512,0.07962133487065633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,768,0.0037902221083641052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,3072,0.008377778033415476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,768,0.004703999807437261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,3072,0.012061333490742577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,2560,0.007106666763623555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,3584,0.019740444090631273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,3072,0.006554666492674086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,8192,0.03462933169470893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,7168,0.010527110762066312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,3072,0.015885333220163982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,3072,0.017068444026841056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,6144,0.033991999096340604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,10240,0.022634666826989915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,5120,0.015595555305480957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,512,0.01881866653760274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,7168,0.012087999946541257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,64,0.004979555391603046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,65536,0.03095910946528117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,6144,2560,0.028224888775083754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,65536,0.4191608958774143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,32,0.004350222233268949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,2048,0.00721955547730128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,5120,0.014067555467287699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,16384,0.023267555567953322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,12288,0.05359733435842726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,16384,0.012859555582205454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,2560,0.008752000000741746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,1536,0.006615110983451207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,768,0.0064702220261096954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,4096,0.014072888427310519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,3584,0.04233155647913615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,2048,0.046446220742331616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,12288,0.01683466633160909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,65536,0.09185510873794556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,12288,0.03921422362327576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,3584,0.01984355515903897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,7168,0.006214222146405115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,128,0.007722666694058313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,5120,0.05526133378346761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,1536,0.009088888764381409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,256,0.005117333183685939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,6144,0.03568089008331299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,512,0.007045333584149678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,2048,0.010607110957304636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,10240,0.0166666673289405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,8192,0.0506586664252811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,64,0.0118568887313207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,16384,0.07597155703438653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,6144,0.012286221815480126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,2048,0.06680888599819608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,4096,0.0064488889442549805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,128,0.008000888758235509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,3584,0.008947555389669206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,65536,0.13573600186242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,2560,0.015297777122921415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,32,0.0052693333062860705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,64,0.0023991111665964127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,65536,0.06645422511630587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,128,0.014754666222466363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,10240,0.013153778182135688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,8192,0.014584889014561972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,5120,0.00889155517021815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,3072,0.007105777661005656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,64,0.006581333362393909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,12288,0.015057777365048727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,65536,0.033561779393090144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,5120,0.031152000029881794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,65536,0.3184666633605957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,7168,0.0336808893415663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,6144,0.1933537721633911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,3584,0.029802666770087347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,512,0.0053324442770746015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,1536,0.005538666827811136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,2560,0.013094222380055321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,7168,0.038093331787321306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,3072,0.12077066633436416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,10240,0.08447555700937907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,768,0.005714666512277391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,12288,0.019156444403860304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,3584,0.021573333276642695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,8192,0.3167857858869764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,128,0.005277333160241445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,128,0.023323555787404377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,10240,0.009034666750166152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,7168,0.011457777685589261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,8192,0.03321333395110236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,5120,0.014187556174066333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,8192,0.3162915441724989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,3072,0.009851555857393477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,768,0.004642666627963384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,2560,1.045447137620714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,2560,0.009726222190592024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,768,0.010470222267839642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,128,0.0037617778612507712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,6144,0.03362577822473314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,8192,0.01814844376511044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,1024,0.0125608891248703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,8192,0.0417484442392985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,16384,0.019296000401178997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,2048,0.008280889027648503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,8192,0.01978133287694719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,8192,0.01128266668981976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,4096,0.02843822373284234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,64,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,8192,0.02738044493728214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,3584,0.012103111379676394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,4096,0.015949333707491558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,3072,0.0069875551594628235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,512,0.008329778081840938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,128,0.0027840000887711844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,4096,0.015025777949227227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,10240,0.006255110932721033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,1024,0.0058755555914507965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,12288,0.05745333433151245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,8192,0.011443555355072021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,12288,0.02156266735659705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,128,0.00720000018676122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,8192,0.00645955569214291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,1536,0.010416000253624385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,1536,0.010318222145239512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,2048,0.007648000286685095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,8192,0.039966222312715315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,256,0.007358222372002072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,10240,0.019117333822780185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,3072,0.0092622223827574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,8192,0.054229332341088184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,64,0.0052133335007561584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,12288,0.05919733312394884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,2048,0.0069875551594628235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,2560,0.005570666657553778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,8192,0.02145066691769494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,6144,0.010306666294733683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,7168,0.021606221795082092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,12288,0.2301128970252143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,1024,0.013380444712109037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,10240,0.03046577837732103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,2048,0.006331555545330048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,16384,0.03338133295377096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,128,0.004639111045334074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,2048,0.008147555920812819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,256,0.008408889174461365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,6144,0.01661777827474806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,3584,0.16643555959065756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,256,0.00721955547730128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,1536,0.12486488289303249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,256,0.011554666691356234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,64,0.007198221981525421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,8192,0.029915554655922785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,3072,0.024458666642506916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,7168,0.0686053302552965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,1536,0.004659555438492033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,1536,0.029029332929187354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,7168,0.010480889015727572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,65536,0.0792177783118354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,12288,0.078329775068495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,4096,0.009646221995353699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,32,0.0033679999825027254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,3072,0.011533333195580376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,12288,0.03360889024204678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,256,0.004578666554556953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,3072,0.006181333214044571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,768,0.003587555554178026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,64,0.005598222215970357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,1536,0.016398222910033334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,6144,0.03014666504330105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,10240,0.021398221453030903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,12288,0.07170222202936809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,6144,0.051474667257732816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,8192,0.045864890019098915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,1024,0.008054222497675154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,1536,0.004583111239804162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,5120,0.0163102216190762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,5120,0.10078133477105035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,128,0.0051893335249688895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,8192,0.012261333564917246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,6144,0.022392888863881428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,768,0.014576888746685453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,65536,0.1793422169155545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,32,0.0025271111064487034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,256,0.002990222225586573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,512,0.016559110747443307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,2560,0.005935110979610019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,65536,0.5937760141160753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,5120,0.029085334804322984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,256,0.008810666700204214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,64,0.0034871109657817413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,8192,65536,0.2897866566975911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,768,0.008107555409272512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,1024,0.07748088571760389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,3072,0.011039999624093374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,64,0.00656444455186526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,3072,0.008466666771305932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,65536,0.1002275546391805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,1536,0.014008000493049622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,128,0.0038044444388813446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,64,0.003519999898142285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,768,0.0054551110499435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,128,0.002611555573013094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,5120,0.07408977879418267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,65536,0.03931822379430135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,256,0.004171555654870139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,16384,0.039480888181262545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,3584,0.006280888699822956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,16384,0.11827288733588324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,12288,0.2941359943813748
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,512,0.005377777748637729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,4096,0.14106310738457575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,3072,0.009062222308582729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,2048,0.00758844448460473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,2048,0.013457777599493662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,768,0.02677244444688161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,256,0.018686221705542672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,1024,0.01386666629049513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,1536,0.019699555304315355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,12288,0.01882755590809716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,1024,0.016720000240537856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,2560,0.008799110849698385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,32,0.0028693332440323303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,3584,0.11953155199686687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,12288,0.04478844337993198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,12288,1.9151662190755208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,128,0.007717333734035492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,10240,128,0.0037875556283526947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,8192,0.16953867011600068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,8192,0.01571111049917009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,1024,0.0075786668393347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,7168,0.048955556419160634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,512,0.008384889198674096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,5120,0.19573244783613417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,128,0.00723555518521203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,12288,0.02507200009293026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,256,3072,0.007098666495747036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,32,0.003632889025741153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,7168,0.18658222092522514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,2048,0.009909333454238044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,12288,0.015584889385435315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,512,2048,0.007365333537260692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,1024,0.09762755367490981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,3584,0.32285332679748535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,512,0.007589333587222629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,7168,0.014817777607176038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,512,0.01054222219520145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,3584,0.09622133440441555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,1536,0.0075128889746136135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,12288,0.25677333937750924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,5120,0.05616355604595608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,2048,0.009500444763236577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,12288,0.027451554934183758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,65536,0.3514711062113444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,4096,0.00906399968597624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,4096,0.018944000204404194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,8192,0.013702222042613559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,64,0.005412444472312927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,16384,0.31739645534091526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,6144,0.018160889546076458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,6144,0.010557333628336588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,10240,0.14791644944085017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,256,0.004513777792453766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,12288,0.07899110847049289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,1024,0.046275556087493896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,4096,0.011634666886594562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,128,16384,0.03397511111365424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,2560,0.006576889091067844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,2048,0.011468444433477191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,512,0.03677155574162801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,768,0.004470222112205294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,32,0.00923288861910502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,3584,0.02861777941385905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,256,0.002985777747299936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,65536,0.008915555146005418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,256,0.006612444503439798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,10240,0.02696711156103346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,3072,0.020986666282018025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,1536,0.0068400001360310455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,256,0.010103999740547603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,1024,0.006235555642180973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,768,0.005894222193294101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,768,0.009411555197503831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,8192,0.1513919962777032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,1536,0.06195022000206841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,7168,0.006445333361625671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,3072,0.010684444672531553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,256,0.008267555799749162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,256,0.00453244439429707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,16384,0.05050666795836555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,2560,0.1156275537278917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,1024,0.868284437391493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,4096,0.029136889510684546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,64,0.008057778080304464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,5120,0.026492445005310908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,12288,0.09994488954544067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,8192,0.014502222339312235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,65536,0.2185466686884562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,2560,0.008199111455016667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,10240,1.51202572716607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,5120,0.009691555466916826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,768,0.0076657773719893555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,32,0.008417777717113495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,512,0.004495999879307217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,2560,0.0064693333374129395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,128,0.007873777714040544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,16384,0.042303111818101674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,6144,0.234186675813463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,64,0.004488888714048597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,10240,0.48923201031155056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,4096,0.011009777585665384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,2048,0.0092257774538464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,6144,0.04123377799987793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,6144,0.029416001505321924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,5120,0.010039110978444418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,4096,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,16384,0.015832000308566622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,8192,0.04615199897024366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,5120,0.02141422198878394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,128,0.0028088889602157804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,5120,0.010337777435779572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,32,0.005967111223273807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,128,0.005016889009210798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,12288,0.04159288936191135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,1024,0.021764443980322942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,1536,0.008613333106040955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,3072,0.034458667039871216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,64,0.005395555661784277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,3072,0.007497777541478475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,32,0.002505777817633417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,128,0.006591111007663939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,65536,14.951483832465279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,6144,0.030838221311569214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,768,0.009469333622190688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,3584,0.01849688920709822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,6144,0.012352888782819113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,6144,0.029636445972654555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,256,0.004260444392760594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,8192,0.028399109840393066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,512,0.006063999815119638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,8192,0.012652444342772165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,12288,0.023596444063716467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,12288,0.042952001094818115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,5120,0.030371556679407757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,1024,0.007731555236710443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,512,0.0034346667428811393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,8192,0.0069617778062820435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,1536,0.0159244438012441
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,128,0.005306666509972678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,12288,0.06637866629494561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,12288,0.026935999592145283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,768,0.00619555554456181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,1536,0.016270221935378183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,12288,0.021008888880411785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,1536,0.0137155552705129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,3072,0.006772444479995304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,32,0.0029164445069101122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,16384,0.30522221989101833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,4096,0.008843555218643611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,8192,0.025814221964942083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,1024,0.03261244297027588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,65536,0.06986844539642334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,16384,0.041165331999460854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,2560,0.01052177780204349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,10240,0.07352888584136963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,2048,0.018183110488785636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,12288,0.5417742199367946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,512,0.005496888938877318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,16384,0.0058648888435628675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,2560,0.01035644445154402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,256,0.005534222142563925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,768,0.061687999301486544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,7168,0.026566222310066223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,64,0.002893333426780171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,6144,0.01014311114947001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,3584,0.012309333516491784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,256,0.01680266691578759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,65536,0.4050773249732123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,256,0.0028239999794297745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,65536,0.23586755328708223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,3584,0.021534222695562575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,3072,0.010144888526863521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,256,0.004800888813204235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,2560,0.006810666786299811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,64,0.0025457777082920074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,8192,0.14468888441721597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,1536,0.05142933461401197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,12288,5.0770649380154085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,65536,0.19130933284759521
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,8192,0.14711111121707493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,128,0.0028640000770489373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,512,0.007111999723646376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,7168,0.021631111701329548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,4096,0.008806222014957005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,3072,0.008745777938101027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,6144,0.025790222816997107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,8192,0.027729777826203242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,6144,0.018017777138286166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,16384,0.02663644486003452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,12288,0.029855112234751385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,7168,0.012178666889667511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,2560,0.006047111004590988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,12288,0.04558755622969734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,7168,0.013415111435784234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,128,0.039000888665517174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,2048,0.013415999710559845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,6144,0.015676443775494892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,8192,0.027132444911532935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,10240,0.07618311378690931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,3584,0.03789688812361823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,12288,0.13534488942888048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,4096,0.02390044430891673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,768,0.003495111233658261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,3072,0.010818666881985135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,10240,0.18953066402011445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,1536,0.007996444072988298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,512,0.016885333591037326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,3072,0.006173333360089197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,10240,0.03607199920548333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,1024,0.006255110932721033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,16384,0.3230675591362847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,6144,0.02713599966631995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,65536,0.49104266696506077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,4096,0.017299556069903903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,4096,0.012375999655988483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,1536,0.009728888670603434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,5120,0.03225244416130914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,128,0.005092444519201915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,3584,0.007430222299363878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,64,0.005568000177542369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,512,0.004946666873163647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,65536,0.23250132136874727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,64,0.005094222310516569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,3072,0.13491377565595838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,2048,0.012157333393891653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,2048,0.006738666858938005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,128,0.0063742221229606206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,8192,0.01185155577129788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,3072,0.006059555543793573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,64,0.005085333353943295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,5120,0.012843555874294706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,10240,0.014448889427714877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,3072,0.022843556271659002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,65536,0.22172889444563124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,3072,0.10482577482859294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,3584,0.05844889084498087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,3584,0.013776000175211163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,7168,0.014853333433469137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,16384,0.022152000003390845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,768,0.008728889127572378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,12288,0.015051555302408008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,3584,0.010684444672531553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,4096,0.017695999807781644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,8192,0.02237066626548767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,4096,0.013455111119482251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,7168,0.010851555400424533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,1536,0.009588444398509132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,128,0.0029093333416514923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,65536,0.032460444503360324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,4096,16384,0.04563733273082309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,3584,0.021831999222437542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,4096,0.007181333170996771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,10240,0.06776533524195354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,8192,0.032772445016437106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,2560,0.016612443659040663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,2048,0.012512889173295764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,256,0.00509688879052798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,128,0.010599999792046018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,5120,0.006870222174459034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,64,0.005688888745175467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,12288,0.029373334513770208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,10240,0.008356444537639618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,32,0.00463733325401942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,128,0.02762577765517765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,768,0.008570666942331526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,1536,0.21168800195058188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,5120,0.01721955504682329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,3072,0.008390222158696916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,512,0.015560888581805758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,7168,0.052856001589033336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,1536,0.0058204444746176405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,3584,0.02360800074206458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,2048,0.018668444620238412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,5120,0.05688889159096611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,768,0.05163644419776069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,3584,0.028351998991436426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,12288,0.01093066649304496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,32,0.0028071111689011254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,64,0.002784888984428512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,3072,0.07813955677880181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,512,0.015031110909250049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,5120,0.011094222466150919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,16384,0.022169777088695105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,4096,0.00813688917292489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,256,0.009903999666372934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,3072,0.006748444504208035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,512,0.003552000141806073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,256,0.012496000362767113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,3584,0.47694842020670575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,6144,0.09502933422724406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,2560,0.008608000146018134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,7168,0.018532445033391316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,256,0.004464000049564573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,7168,0.04150399896833632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,7168,0.12215110990736221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,32,0.002728888971938027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,16384,0.2781617906358507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,3072,0.007725333174069722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,512,0.007270221908887227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,128,0.004086222085687849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,12288,0.04433688852522108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,12288,0.0322693321439955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,7168,0.013044444223244986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,65536,4.784038967556424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,2048,0.008808000220192803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,2048,0.011448889142937131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,10240,0.039289779133266874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,4096,0.01573866605758667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,3072,0.006797333144479328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,32,0.002903999967707528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,3072,0.009571555587980483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,2560,0.009209777745935652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,12288,0.0399271117316352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,6144,0.010294222169452244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,256,0.01938488913906945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,5120,0.4859617551167806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,64,0.01805688937505086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,3584,0.06723911232418485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,16384,0.06285599867502849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,12288,0.31957512431674534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,65536,0.18971910741594103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,5120,0.014263111684057446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,64,0.003744889050722122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,2560,0.008866666919655269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,5120,0.01904533306757609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,128,0.013386666774749756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,12288,0.01851466629240248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,65536,0.41173066033257383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,12288,0.05442844496832954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,3584,0.015498666299713982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,16384,0.050717333952585854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,65536,0.4105822245279948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,8192,0.20204533471001518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,128,0.008774222599135505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,8192,0.020799999435742695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,3072,0.013243555194801755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,6144,0.03204266561402215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,8192,0.06779822376039293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,2560,0.03143289022975498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,256,0.008432000047630733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,64,0.005520888914664586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,7168,0.022848889231681824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,1536,0.00721688899728987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,3072,0.006796444455782573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,1536,0.010803555448849997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,7168,0.026348445150587294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,4096,0.010987555815113915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,7168,0.05637511279847887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,7168,0.049103998475604586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,6144,0.014607111612955729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,12288,0.02437066700723436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,256,3072,0.0074684446056683855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,12288,0.016556445095274184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,2560,0.009715555442704095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,5120,0.015791111522250705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,2048,0.00849777791235182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,12288,0.015217777755525378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,768,0.011284444067213269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,128,0.004592888885074192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,1024,0.01864977843231625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,768,0.006436444405052397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,65536,1.7285422219170465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,512,0.006549333532651265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,10240,0.02090311050415039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,1536,0.005783111270931032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,32,0.0063688887490166565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,4096,0.054654220740000405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,6144,0.37772178649902344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,5120,0.12222222487131755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,16384,0.059078223175472684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,256,0.004398222184843487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,768,0.01830222209294637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,6144,0.019883554842736986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,2048,0.010445333189434476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,65536,0.42011022567749023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,16384,0.29568711916605633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,2048,0.013720889058378009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,8192,0.041024890210893415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,64,0.0026355555488003623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,6144,0.01089511149459415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,16384,0.2796853383382161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,10240,0.03281777766015794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,1024,0.004087111188305749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,512,0.009372444616423713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,4096,0.0063546668324205614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,10240,0.07256888680987887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,5120,0.018928888771269057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,32,0.005017777697907554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,512,0.5833368831210667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,128,0.015947555502255756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,65536,0.08144088586171468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,5120,0.027307555079460144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,768,0.007284444239404466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,3584,0.8794622421264648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,2560,0.011535999675591787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,10240,5120,0.06174933248096042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,512,0.025125333004527624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,32,0.003481777591837777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,128,0.0069226668112807805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,6144,0.026535110341178045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,5120,0.023216888308525085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,6144,0.016566221912701923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,1536,0.055861334005991616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,6144,0.006502222269773483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,5120,0.012668444050682915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,8192,0.04483022292455038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,512,0.0070364442136552595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,6144,0.678427537282308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,3584,0.03263555632697211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,2560,0.008952000074916417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,128,0.005066666752099991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,6144,0.03394488824738397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,512,0.03046577837732103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,64,0.006549333532651265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,10240,0.04648177822430929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,12288,0.019419555862744648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,5120,0.24741066826714408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,256,0.0046826667255825466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,12288,0.023741333021057978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,16384,0.03380799955791897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,65536,1.891329871283637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,3584,0.018056000272432964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,4096,0.054321779145134814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,5120,0.00701155596309238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,128,0.014004444082578024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,8192,0.37138933605617946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,3584,0.03121244576242235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,7168,0.02364088926050398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,32,0.005559111220969095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,4096,0.006269333263238271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,4096,0.008760888543393876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,65536,0.02751200066672431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,1024,0.011955555942323474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,3584,0.009015111459626092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,16384,0.07047910822762384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1024,256,0.007215110792054071
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,32,0.006861333217885759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,6144,0.0058400001790788435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,2048,0.007144889069928064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,3072,0.011735999749766456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,4096,0.35565688874986434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,2560,0.013111111190583972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,1536,0.13328266143798828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,5120,0.3098826673295763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,256,0.005114666703674528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,12288,0.029101332028706867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,768,0.04935555656750997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,5120,0.012280888855457306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,4096,0.023749333288934495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,1536,0.012338666452301873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,12288,0.01930488811598884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,8192,0.10223733054267035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,2048,0.0064186664919058485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,5120,0.03631022241380479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,4096,0.00851022203763326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,10240,0.04551466637187534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,3072,0.09022488858964707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3072,2560,0.012807110945383707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,10240,0.022259554929203455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,2048,0.08041955365075006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,6144,0.025225778420766194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,512,0.13580444124009874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,128,0.004052444464630551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,10240,0.17676799827151826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,3072,0.006559111177921295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,4096,0.014688889185587565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,768,0.011944000091817645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,256,0.0063546668324205614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,16384,0.05099466774198744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,5120,0.01034222212102678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,512,0.013367999758985309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,3584,0.06910933388604058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,12288,0.11633422639634873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,65536,0.6883324517144097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,7168,0.06492711438073052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,4096,0.030113776524861652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,10240,0.014135999812020196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1536,2048,0.006762666834725275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,1024,0.005984888722499211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,6144,0.04665866825315687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,12288,3584,0.016883555385801528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,256,0.003104000041882197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,12288,1.064318233066135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,10240,0.023388443721665278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,256,0.06919466786914401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,8192,0.24388978216383192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,16384,0.008408889174461365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,32,0.0031031111462248694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2048,7168,0.015030221806632148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,1024,0.010464000205198923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,7168,0.18854933314853242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,256,0.005314666777849197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,6144,0.014173333843549093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,1024,0.008072888685597314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,3584,0.006635555376609166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,5120,0.22581243515014648
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3584,256,0.004648888690604104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,1024,0.012000888586044312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,4096,0.01925155520439148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,512,0.021355556117163763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,7168,0.0476835568745931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,6144,3584,0.014456000592973499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,8192,0.0902728901969062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,128,0.0026506665680143568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,65536,0.32970576816134983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,256,0.0042204442951414315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,4096,0.008089777496125963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,1536,0.008192888564533657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,64,0.00291377781993813
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,2048,0.0111191107167138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,8192,0.11581155988905166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,2560,0.021314667330847845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,3584,0.00962488849957784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,2560,0.009384888741705153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,128,0.004480888860093223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,6144,0.15519821643829346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,512,0.005557333429654439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,768,0.11880621645185684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,1024,0.008054222497675154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,2048,0.007300444775157505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,512,2048,0.005881777654091517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,128,0.004012444367011388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,512,0.0031315556002987754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,3584,0.007505777809354994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,512,0.005069333232111401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,3072,0.007726222276687622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,64,0.0031439999325407874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,3584,0.008998221821255153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,8192,0.1491022242440118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,1536,0.009024889104896123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,2560,0.009025777379671732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,8192,0.13067111704084608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,12288,2.4888409508599176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,10240,0.037110222710503474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,128,0.006244444598754247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,128,0.004357333398527569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,32,0.008932444784376357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,256,5120,0.011090666883521609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,768,0.015711999601787992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,8192,0.2495457861158583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,12288,0.030047999487982854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,10240,0.01687022215790219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,64,0.0053084443012873335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,5120,0.012549333274364471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,12288,3072,0.022450667288568284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,3584,0.01676977839734819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,2560,0.014212444424629211
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,64,0.008617777791288164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,4096,0.01572711103492313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,16384,0.031389332479900785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,3072,0.2325404485066732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,256,0.0028986665937635633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,12288,0.008617777791288164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,10240,0.011356444822417365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,2560,0.005997333261701796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,10240,0.027082666754722595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,2560,0.014836443795098199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,32,4096,0.009712888962692684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,16384,0.04269511169857449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,512,0.005292444593376583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,4096,0.010341333018408881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,1536,0.00776800016562144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,2560,0.01882488860024346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,3072,0.007394666473070781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,8192,0.01938222183121575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,4096,0.005664000080691443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,65536,0.06435288985570271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,512,0.01831466621822781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,2048,0.013571555415789286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,768,0.00828888929552502
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,10240,0.007170666423108842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,10240,0.017786666750907898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,3584,0.07805244127909343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,5120,16384,0.03164800008138021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,10240,0.055884447362687856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,65536,1.7095022201538086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,32,0.0033928888539473214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,1024,0.00776444458299213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,128,0.003039111072818438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,1536,0.05249333381652832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,768,0.019038221902317472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,3072,0.017703110973040264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,10240,0.03699466586112976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,10240,0.03204088740878635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,512,0.003765333443880081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,12288,0.02805866797765096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,3072,0.014715555641386243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,16384,0.05594578054216173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,8192,0.023915555742051866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,8192,0.1681333382924398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,7168,0.022984888818528917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,5120,0.043707556194729276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,64,0.019333332777023315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,6144,0.01623911162217458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,16384,0.022247110803922016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,768,0.005396444350481033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,128,0.005335999859703912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,1536,0.005961777849329843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,10240,0.016563556260532804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,256,0.003910222401221593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,768,0.008758222063382467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,256,0.0032311110860771606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,16384,1.2659022013346355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,32,0.0032088889016045462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,1536,0.011319110790888468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,2048,0.22775912284851074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,65536,0.07051377826266818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,4096,0.07543200254440308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,3584,0.0229120006163915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,7168,0.006227555374304454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,4096,0.01850044396188524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,2560,0.022031999296612207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,5120,0.034879999028311834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,1024,0.009738666315873465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,3584,0.0064479998416370815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,2048,0.006093333164850871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,768,0.008912000391218398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,512,0.009717333647939894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,4096,0.025238222546047632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,5120,0.08467911349402533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,2560,0.01587733295228746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,16384,0.2865777810414632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,768,0.006074666563007567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,1024,0.006424888968467712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,16384,0.13881066110399035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,6144,0.025198222862349615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,3584,0.23941156599256727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,1536,0.017786666750907898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,5120,0.5472284422980415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,2560,0.01572088897228241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,3072,0.007107555866241455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,512,0.005498666730191972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,6144,0.023736889163653057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,4096,0.007605333295133378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,5120,0.09988444381290013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,7168,0.3437973393334283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,7168,0.016155555844306946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,65536,6.74053955078125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,3584,0.007352888584136963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,3072,0.05385066734419929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,5120,0.06992977857589722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,2560,0.011035555766688453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,6144,0.01928622192806668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,256,0.007798222204049428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,3072,0.02715377840730879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,256,0.0030417777597904205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,1536,0.0074471111098925276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,3072,0.1026817758878072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,2048,1.6858569251166449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,16384,0.05940622091293335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,4096,0.011311111350854238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,256,0.004864888886610667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,1536,0.09395911296208699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,10240,0.03358666764365302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,1024,0.005668444352017508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,12288,0.02678844498263465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,3072,0.044050667021009654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,64,0.008035555481910706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,65536,3.5769271850585938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,12288,0.11328532960679795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,6144,0.014508444401952954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,1536,0.009762666291660732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,128,0.010224888722101847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,16384,0.09687466753853692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,7168,0.023565333750512864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,5120,0.030429333448410034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,3072,0.008240888516108194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,10240,0.006918222126033571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,512,0.017044444878896076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,512,0.0470142232047187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,6144,0.01016977760526869
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,10240,0.018555555078718398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,256,0.010556444525718689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,4096,0.21429599655999076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,1024,0.05662844578425089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,16384,768,0.012728888955381183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,512,0.005544888890451855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,256,0.005281777845488654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,8192,0.08659288618299697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,768,0.005605333381228977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,12288,0.025951999757024977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,3584,0.012261333564917246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,256,0.002895111011134254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,256,0.0034737777378824023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,16384,0.041725334193971425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,3072,0.020571556356218126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,65536,0.3120337857140435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,8192,0.029701332251230877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,4096,0.010377777947319878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,4096,0.010953777366214328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,2560,0.013857777747843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,768,0.01756622228357527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,1024,0.0058542220956749385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,8192,0.014477334088749357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,512,0.005430222385459476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,5120,0.039275556802749634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,4096,6.619142320421006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,5120,1.1044996049669054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,4096,0.05462222297986349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,8192,0.01538222200340695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,7168,0.011340444286664328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,6144,0.01993777851263682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,5120,0.015648888217078317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,128,0.009520000053776635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,768,0.0038977778620190094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,10240,0.014408888088332282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,4096,0.06118577718734741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,10240,0.00962044464217292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,128,0.0028693332440323303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,5120,0.006688000013430913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,128,2048,0.007113777928882175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,4096,0.008719999757077958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,16384,0.02072266737620036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,12288,0.0628017783164978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,3072,0.007038222418891058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,6144,0.024287111229366724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,4096,0.01998933321899838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,16384,0.006679111056857639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,1024,0.007672888537247975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,1024,0.005763555566469829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,16384,0.08780444330639309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,8192,0.01129244433508979
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,2048,0.012412444584899478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,10240,256,0.015488889482286243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,5120,0.03228800164328681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,8192,0.11352445019616021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,512,0.0038693332009845306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,1024,0.010020444790522257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,2560,0.07380000087949964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,256,0.0053839998112784485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,128,0.006955555743641323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,2560,0.033795555432637535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,2560,0.007034666836261749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,12288,0.018560000591807894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,5120,0.06237688991758558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,3584,0.06362222300635444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,10240,0.020440000626775954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,64,0.005370666583379109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,7168,0.015251555376582675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,10240,0.06389777527915107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,256,0.006219555520349079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,8192,0.02527911133236355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,1536,0.017218665944205392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,16384,0.04845688740412394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,5120,0.030062221818500098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,2560,0.015292444162898593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,5120,0.04144710964626736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,2048,0.059843553437127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,2560,0.015055111712879606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,256,0.004299555387761858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,512,0.00498933345079422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,768,0.00546488869521353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,3584,0.014283556077215405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,1024,0.004034666551484002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,2048,0.013585777746306526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,3584,0.01977688901954227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,64,0.005205333232879639
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,768,0.004882666799757215
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,5120,0.04067466656366984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,12288,0.3977368937598334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,6144,0.012332444389661154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,6144,12288,0.3956853283776178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,512,0.006162666612201267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,5120,0.019467555814319186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,16384,0.024311999479929607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,5120,0.03311199943224589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,1536,0.00675733346078131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,8192,0.024149333437283833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,1536,0.009475555684831407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,128,0.008609777523411645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,5120,0.013162666724787818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,512,0.00370755543311437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,64,0.0027457777824666765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,768,0.006582222051090664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,3072,0.00847466703918245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,2048,0.01018488903840383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,10240,0.01348977784315745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,8192,0.018228444788191054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,1536,0.021362667282422382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,4096,0.7805217636956109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,512,0.004228444563017951
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,768,0.012671999633312225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,8192,0.03424888849258423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,768,0.004332444320122401
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,32,0.00544088871942626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,3584,0.009081777599122789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,768,0.013677332964208392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,16384,0.1052008867263794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,512,0.006226666685607698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,7168,0.020992888344658747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,7168,0.026896889011065166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,16384,2.4630194769965277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,4096,0.020727111233605277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,2048,0.014932443698247274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,4096,0.02549422283967336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,128,0.004388444539573458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,2048,0.03760622276200189
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,7168,0.007094221810499827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,512,0.030279109875361126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,2560,0.02472266720400916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,2048,0.007686222592989604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,256,0.006103111224042044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,1536,0.005513777749405966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,3584,0.024712888730896845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,12288,0.24358577198452422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,65536,0.06304533614052667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,1536,0.014361777239375643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,7168,0.021116443806224402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,65536,0.0074879998962084455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,12288,0.007832888927724626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,8192,0.11640355322096084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,65536,0.34772443771362305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,768,0.005763555566469829
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,64,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,3584,0.016619554824299283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,3072,0.011223110887739392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,1024,0.004263111286693149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,4096,0.022579555710156757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,5120,0.00848355558183458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,4096,0.020971554848882887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,128,0.0032906666811969546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,1024,0.0037973332736227247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,5120,0.04599644409285652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,8192,0.7211075358920627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,2560,0.006831999868154526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,16384,0.09047200282414754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,12288,0.054677334096696645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,10240,0.015192000402344598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,1536,0.009864889085292816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,12288,0.026800000005298193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,1024,0.009575111170609793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,128,0.005241777747869492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,256,0.0038968887594011095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,12288,0.05529866615931193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,512,0.004880889008442561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,65536,16384,0.27577421400282115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,3584,0.014389332797792224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,12288,0.059691554970211454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,10240,0.018295110927687753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,128,0.005859555469618903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,16384,0.017198221551047433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,128,0.015589333242840238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,10240,12288,0.038203554020987615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,8192,0.04956000049908956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,5120,0.09406666623221503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,1536,0.010282666318946415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,65536,0.4637120034959581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,3072,0.00740799970097012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,128,0.02535555594497257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,8192,0.028276443481445312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,128,0.004256000121434529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,7168,0.12331288390689427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,6144,0.00664533343580034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,4096,0.024078221784697637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,4096,0.014090667168299357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,768,0.006491555521885554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,10240,0.039108445247014366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,32,0.008731555607583787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,4096,0.3631262249416775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,16384,0.04115110966894362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,5120,2048,0.01127911110719045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,10240,0.03720088799794515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,3072,0.01368799971209632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,512,0.0032462223122517266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,768,0.008716444174448649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,12288,0.006385777973466449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,4096,0.006215111249023014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,16384,0.019712888532214694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1024,768,0.003597333199448056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,65536,0.023216888308525085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,7168,0.6469884448581272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,128,0.0041519999504089355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,7168,0.08216977781719632
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,3072,0.011549333731333414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,4096,0.02624711063173082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,7168,0.008608000146018134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,256,0.0037324445115195382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,512,0.007794666621420119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,4096,0.007867555651399825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,6144,0.08340266678068374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,7168,0.03159022331237793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,64,0.006828444285525217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2048,10240,0.015180443723996481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,4096,0.6431671248541938
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,1536,0.012331555287043253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,10240,0.04115733173158433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,5120,0.022111111217074927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,64,0.0025600000388092464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,4096,0.031870222753948636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,1024,0.0075662218862109715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,1024,0.02293333411216736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,10240,0.036446223656336464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,7168,0.017495999733606975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,2560,0.010647999743620554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,2560,0.009490666290124258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,256,0.0028862222615215513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,256,0.003060444361633725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,512,0.007408888803588019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,16384,0.021762667430771723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,7168,0.03007733490731981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,2048,0.02370400064521366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,6144,0.027481777800454035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,3584,0.018233777748213876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,8192,0.01149511088927587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,8192,0.011952000359694162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,768,0.01721333298418257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,128,0.0052391112678580815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3072,3584,0.009879111415810054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,32,0.0028391112056043413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,65536,0.6759982109069824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,8192,0.006599111275540457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,3072,0.015092444088723926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,12288,2560,0.013696889082590738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,10240,0.014875556031862894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2048,10240,0.32851823170979816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,128,0.053080889913770885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,1024,0.07573599947823419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,16384,0.02384977704948849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,512,0.007887110941939885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,768,0.006002666635645761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,1024,0.006020444548792309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,512,0.0037191112836201987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,512,0.014372444815105863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,128,0.004309333446953031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,8192,0.03163199954562717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,3072,0.008224888808197444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,8192,0.02829155657026503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3584,3072,0.016507556041081745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,3584,0.006168889088763132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,5120,0.058396445380316846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,32,0.003918222255176968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,8192,0.03875022133191427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,10240,0.028104000621371802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,4096,0.18263199594285753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,3072,0.11750933859083389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,512,0.0032053333189752367
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,2048,0.006888000087605582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,768,65536,0.20522844791412354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,3584,0.00830666638082928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,8192,0.22467022471957734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,4096,0.025615110993385315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,3072,0.016608888904253643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,65536,0.24315643310546875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,16384,0.08723111285103692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,768,0.005531555662552516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,256,0.0037626665499475268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,6144,256,0.006375999914275275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,16384,0.009520000053776635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,128,0.002798222212327851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,12288,0.04040177663167318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,12288,0.03903999924659729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,2560,0.0131217779384719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,32,0.0028986665937635633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,8192,0.012169777519173093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,3584,0.008098666866620382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,12288,0.01824355622132619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,1536,0.00489066665371259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,16384,0.012903999951150684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,10240,0.01718577742576599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,1024,0.17219466633266875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,7168,0.012098666694429187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,1536,0.01038577738735411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,2560,0.009000000026490953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,4096,0.010997333460383944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,128,0.004420444369316101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,4096,0.007701333198282454
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,32,0.004257777912749184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,65536,0.0668213301234775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,3072,0.34788889355129665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,128,0.006059555543793573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,65536,0.4275048838721381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,16384,0.10259733597437541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,512,0.006291555447710886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,3072,0.00904977818330129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,1024,0.006461333483457565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,4096,0.018619555566045973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,12288,0.05749066670735677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,512,0.005803555664088991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,2048,0.005230222311284807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,1024,0.011424000064531961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,256,0.0064080001579390625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,1536,0.005707555760939916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,10240,0.04483022292455038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,3072,0.010210666391584609
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,12288,0.08969333436754014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,128,0.005342222336265776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,3072,0.02223644488387638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,16384,0.02498577700720893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,65536,1.5453191333346898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1536,2560,0.009008888569143083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,6144,0.013414222333166333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,1536,0.006508444332414203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,1536,0.016190222567982145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,3584,0.008105778031879002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,128,0.0052871112194326185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,2560,0.006741333338949416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,6144,0.01313688854376475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,3072,0.06005066633224487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,512,0.004127999974621667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,16384,0.05153600043720669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,65536,0.0310791101720598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,512,0.007841777470376756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,512,0.0032506665835777917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,32,0.005164444446563721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,2560,0.006531555619504716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,256,0.007962666451931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,12288,0.35538045565287274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,32,0.003035555490189128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,65536,0.23504355218675402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,16384,0.07791999975840251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,3072,0.0692888895670573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,8192,0.03749955693880717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,2048,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,5120,0.022670222653283015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,2560,0.008560000194443597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1536,7168,0.011227555572986603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,768,0.013428444663683573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,5120,0.008596444295512306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,8192,0.7403990957472059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,10240,0.01757866640885671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,5120,256,0.00535999983549118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,6144,0.025083555115593806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,10240,0.008430221842394935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,3584,0.01184622198343277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,6144,0.012334222594896952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,3584,0.01630577776167128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,7168,0.011125333607196808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,768,0.005646222167544895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,6144,0.011797332929240333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,256,0.0038480001191298165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,6144,0.019545776976479422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,6144,0.030163556337356567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,128,0.0038746665749284956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,4096,0.010623111493057676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,3584,0.01889866590499878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,7168,0.006191111273235745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,256,0.004316444612211651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,8192,0.006793777975771163
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,10240,0.042600890000661217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,4096,0.008552889029184977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,8192,0.023078221413824294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,10240,512,0.005552000055710475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,8192,0.014564444621404013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,128,0.005312888986534542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,3584,0.01163999984661738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,256,0.015708444847000968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,64,0.0053119998839166425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,128,0.004257777912749184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,6144,0.0144195556640625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,1536,0.19466488891177705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,16384,16384,0.13891377713945177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,256,0.007914666500356462
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,512,256,0.003974222060706881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,16384,0.10579733053843181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,128,0.004242666479614046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,6144,16384,0.05535377727614509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,12288,0.0425600012143453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,12288,0.07468622260623507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,3072,0.016206221448050607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,768,2560,0.01899111105336083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,7168,0.026159110996458266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,16384,0.03421155611673991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,1024,0.39801067776150173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,12288,0.3787919945187039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,768,0.006216889040337668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,7168,0.014392889208263822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,32,0.008611555728647444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,64,0.002854222224818336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,2560,0.023931556277804907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,1536,0.012582222620646158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,1024,0.020795555578337777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,65536,1536,0.03243288728925917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,64,0.005118222286303838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,16384,0.06033244397905138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,768,0.005930666708283954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,8192,0.04372444417741564
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1024,10240,0.03344800074895223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,7168,0.040215111441082425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,7168,0.030234667989942763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3584,512,0.009519110951158736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,1024,0.004020444634887907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,5120,0.03286044465170966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,3584,0.017394666870435078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,65536,0.02602311140961117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,128,0.016081778539551627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,16384,0.04914933443069458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,3072,0.008437333007653555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,6144,0.19023111131456164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,1536,0.12822043895721436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,1024,0.00850044439236323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,1024,0.005668444352017508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,12288,0.016553777787420485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,64,0.008446222378147973
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,4096,768,0.005299555758635203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,16384,0.019923556182119582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,8192,0.03387377659479777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,5120,3072,0.053973333703147046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,4096,0.009478222164842818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,256,0.00870311094654931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,64,0.005418666534953647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,16384,0.022295110755496558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,3584,0.007544000115659501
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,1536,0.004769777672158347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,2560,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,64,0.006414222220579783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,2048,0.005341333233647876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,7168,0.13668978214263916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,8192,0.054213331805335156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,128,0.004492444296677907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,1536,0.01697688963678148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,4096,0.015959110524919298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,5120,0.009096889032257928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,6144,0.01147911118136512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,12288,256,0.07496888769997491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,2560,0.04155644443300035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,2048,0.015896000795894198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,256,0.003452444449067116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,7168,0.010748444332016839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,16384,0.05937600135803223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,2048,0.07574844360351562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,6144,0.008029333419269985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,128,0.012237333589129977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,16384,2560,0.0523173345459832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,1024,0.01458666721979777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,3584,0.2085813283920288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,1024,65536,0.06599822309282091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,3584,0.015392000476519266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,4096,0.00831288927131229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,12288,0.023807111713621352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,4096,0.09234844313727485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,64,0.00869599978129069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,5120,0.01388711151149538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,2560,0.013304000099500021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,128,0.0025920000755124623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,3072,0.007360888852013483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,4096,0.015768888923856948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,768,0.003653333418899112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,3072,0.008419555922349295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,8192,0.030508445368872747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,2048,0.08071911334991455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,32,0.002624888800912433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,3584,0.006844444407357111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,2560,0.01737422247727712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,4096,0.008343111309740279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,64,0.011925333076053195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,6144,0.02464977734618717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,256,0.005671111245950063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,16384,0.012004444168673621
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,2560,0.013036444783210754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,512,0.0031724443866146933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,16384,0.5306746694776747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,2560,0.006119110931952794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,768,0.01573777695496877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,8192,0.01723377737734053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,256,0.008079999850855933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,2560,0.11539199617173936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,128,0.006413333531883027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,512,0.005193777796294955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,3072,0.273510217666626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,4096,0.017307556337780423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,8192,0.03525422347916497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,64,0.007854222423500484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,2048,0.007726222276687622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,12288,0.01722311145729489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,10240,0.00655733338660664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,512,0.005163555757866966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,7168,0.011380444798204633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,6144,0.034128000338872276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,512,0.010124444133705562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,4096,0.007682667010360294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,3072,0.006516444600290722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,65536,1024,0.22346311145358613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,1024,0.0064186664919058485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,3072,0.008576889004972246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,3584,0.011207111179828644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,65536,0.06600089205635919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,3584,0.3206915590498182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,5120,0.022478222846984863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,65536,0.09091999795701768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,1024,0.01755288905567593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,32,0.0053004444473319584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,10240,0.006933333145247565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,7168,0.026924444569481745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,5120,0.01958311100800832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,512,0.006846222198671765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,32,0.0029484445436133277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,1024,0.06265688604778714
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,5120,0.03343111276626587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,2560,0.00906133320596483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,5120,0.006642666541867786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,65536,0.21680977609422472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,32,0.003616888903909259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,512,0.004519111166397731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,2560,0.18591554959615073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,768,0.003584889074166616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,2048,0.03713066710366143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,2048,0.006878222028414409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,5120,0.011954666839705573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,128,0.007281777759393056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,3072,0.011108444796668159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,6144,0.018516444497638278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,512,0.0391937792301178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,1536,0.011936888926559024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,12288,2.09262572394477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,512,256,0.0051893335249688895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,768,0.005607111172543631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,512,0.0046942221621672315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,5120,0.011234666738245221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,256,0.0047493332790003884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,256,0.004451555510361989
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,3584,0.008739555875460306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,2048,0.046281778150134616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,16384,0.10459466775258382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,4096,0.007600888609886169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,1536,0.00832799987660514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,1024,0.04874844352404276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,2048,0.013603555659453074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,3072,0.06275022029876709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,16384,0.010483555495738983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,1024,0.00943999985853831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,64,0.0051893335249688895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,1536,0.014907555447684394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1024,256,0.005542222410440445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,3072,0.007167110840479533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,2048,0.006671999891599019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,32,0.0030906666070222855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,768,0.0052373334765434265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,3072,0.01608088943693373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,1536,0.00980799976322386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,2560,0.007183111376232571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,5120,1024,0.006559111177921295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,3584,0.034303999609417386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,3072,0.07013955381181505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,7168,0.014139556222491793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,3584,0.008016889293988546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,16384,0.05968266725540161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,256,0.008513777620262569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,12288,0.12176622284783258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,768,0.005879999862776862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,65536,0.21000177330440947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,2560,0.0063928887248039246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,2560,0.012539555629094442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,768,0.00435288871328036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,7168,0.02456444501876831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,6144,0.031133333841959637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,1024,0.005882666756709416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,5120,0.038555555873446994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,4096,0.5436951319376627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,1024,0.006124444305896759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,768,0.010783111055692038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3584,65536,3.5469199286566844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,10240,2048,0.3038960032992893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,3072,0.18973333305782744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,10240,0.016007110476493835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,2048,0.009304000271691216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,64,0.002855111120475663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,32,0.005883555445406172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,1536,0.006020444548792309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,5120,0.009944000177913243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,768,0.005312888986534542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,10240,0.16445333427853054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,2560,0.008166222108734978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,16384,0.12583733929528132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,3584,0.011446221835083432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,128,0.003169777699642711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,64,0.06298577785491943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,768,0.0037493333220481873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,8192,0.07029422124226888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,7168,0.04670666654904684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,2048,0.00608711110221015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,128,0.006839111033413146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,6144,0.00997155573632982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,3584,0.4480195575290256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,768,0.049476444721221924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,10240,0.027492443720499676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,64,0.007140444384680853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,128,8192,0.03694222370783488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,8192,0.012149333126015134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,7168,0.19247378243340385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,256,0.0028515555378463534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,1536,0.00904177791542477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,65536,0.6136239899529351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,4096,0.009399111072222391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,16384,0.025966222087542217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,512,0.009443555441167619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,1536,0.014890667464998035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,7168,0.03744622071584066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,128,0.004087111188305749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,1536,0.011936888926559024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,64,0.005099555684460534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,8192,0.012034667034943899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,5120,0.02644888891114129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,6144,0.01111288865407308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,768,0.003616888903909259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,65536,0.24659556812710234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,12288,0.043326222234302096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,4096,0.038471112648646034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,65536,0.1501973337597317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,16384,0.05077955457899305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,2048,0.008042666647169325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,512,0.025143999192449782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,7168,0.3430942164527045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,2048,0.00794044468137953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,3072,0.034006221426857844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,2048,0.286260445912679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,6144,0.010507555471526252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,768,0.031952887773513794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,8192,0.022437334060668945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,256,0.0032915555768542816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,1536,0.006723555425802867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,256,0.004267555558019214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,768,0.07975644535488553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,7168,0.026742221580611333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,7168,0.009958222508430481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,2048,0.006689777804745569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,7168,0.36406222979227704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,10240,0.017833777599864535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,3072,0.018579555882347953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,8192,0.01811733345190684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,3072,0.029657777812745836
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,3584,0.014170666535695394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,2560,0.007732444339328342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,5120,0.05827022261089749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,3584,0.01457155578666263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,128,0.0037582222786214617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,2048,0.00582755563987626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,7168,0.023552000522613525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,1536,0.006143110907740063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,6144,0.012870222330093384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,16384,0.048334221045176186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,64,0.006022222340106964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,12288,0.0717031094763014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,2560,0.014476444986131458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,7168,0.03647911217477586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,512,0.004300444490379757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,10240,0.02702666653527154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,65536,0.07275733020570543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,7168,0.01293511109219657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,512,0.005994666781690385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,7168,0.04208977686034309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,10240,0.013576000101036496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,5120,512,0.005809777726729711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,8192,0.03842755489879184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,5120,0.012087999946541257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,16384,0.2822035683525933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,3584,0.01295111080010732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,2560,0.006079999936951532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,768,0.0033822223130199644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,3584,0.1668142212761773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,7168,0.014610666367742749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,2048,0.011580444044537015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,2048,0.0075137780772315125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1024,32,0.006503110958470239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,1536,0.00489066665371259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,10240,0.009517333573765224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,12288,0.015271999769740634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,10240,0.03793688946300083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,512,0.0038399998512532976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,7168,0.23364623387654623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,65536,0.013569778038395775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,2048,0.006974221931563483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,64,0.2317119969262017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,5120,0.009240888886981541
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,64,0.011578666667143503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,65536,0.07506844732496473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,1024,0.06195377641254001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,10240,10240,0.03557777735922072
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,16384,0.21566489007737902
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,4096,0.009685333404276106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,16384,0.02012888921631707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,256,0.0370435549153222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,256,0.007095110913117726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,256,0.004751111070315043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,2048,0.018263111511866253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,7168,0.012846222354306115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,4096,0.018928888771269057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,16384,0.103928890493181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,10240,0.17757866117689344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,512,0.005509333478079901
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,10240,0.3007395532396105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,3584,0.0074942219588491656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,2560,0.005997333261701796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,1536,0.03847022189034356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,256,0.005190222213665645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,2048,0.007222221957312689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,3584,0.006913777854707506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,1024,0.003939555750952827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,12288,0.06237777736451891
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,3072,0.016308445069524977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,3584,0.012243555651770698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,4096,0.01975555552376641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,256,3072,0.007916444705592262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,3584,0.008564444051848518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,6144,0.026384888423813715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,16384,0.011689778003427716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,16384,0.5083262125651041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,10240,0.02036977807680766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,1536,0.009090666969617208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,3072,0.006612444503439798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,3072,0.029079112741682265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,12288,0.01625155574745602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,8192,0.009840889109505547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,10240,512,0.012558221817016602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,12288,0.06785777542326185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,128,0.0282008879714542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,16384,0.020350222786267597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,128,0.006225777582989798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,12288,0.014967110421922473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,16384,0.02682133350107405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,512,0.005340444544951121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,64,0.005131555514203178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,3072,0.007655999726719326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,3072,0.006519111080302133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,2560,0.01590399940808614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,65536,0.3886524571312799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,7168,0.006584888945023219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,128,128,0.0053173332578606075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,512,0.004490666505363253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,5120,0.010672888822025724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,10240,0.016173332929611206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,7168,0.006550222221348021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,6144,0.03928088810708787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,1536,0.009229333036475712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,768,0.0034462221794658234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,512,6144,0.007510222494602203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,6144,0.019016888406541612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,1024,0.007993777592976889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,1536,0.013454222016864352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,256,0.008481777376598781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,12288,0.02039377722475264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,3584,0.014058666096793281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,6144,0.02693422304259406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,65536,1.5166631274753148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,2560,0.008608889248636033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,4096,0.01406133340464698
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,64,0.009436444275909001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,8192,0.01587466730011834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,256,0.0033848887930313745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,3584,0.22546399964226616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,1536,0.021596444977654353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,7168,256,0.03408177693684896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,64,0.0025902222841978073
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,3584,0.006799110935793982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,16384,0.029489778810077246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,512,7168,0.010942222343550788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,128,16384,0.2067884471681383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,10240,0.01663644446267022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,2048,0.019509333703253005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,256,0.005789333333571752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,256,0.00423555572827657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,5120,0.03087644444571601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,4096,3584,0.19953866799672446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,12288,0.04404977626270718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,16384,0.020280000236299302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,12288,0.02606133288807339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,10240,0.020999999509917364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,3584,0.02122755514250861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,3584,0.007366221812036302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,512,0.00886577781703737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,16384,0.3903795613182916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,4096,0.13987199465433756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,10240,2560,0.014239999983045789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,65536,0.4760817951626248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,65536,0.18518933984968397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,3072,0.10676444239086574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,8192,2048,0.009766221874290043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,7168,0.03198755449718899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,8192,0.03968800107638041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,128,0.0053431110249625305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,7168,0.00740444411834081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,65536,0.02068711154990726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,4096,4096,0.018214222457673814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,128,0.09386489126417373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,512,0.007344000041484833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,3584,0.05159822106361389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,65536,0.193122665087382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,7168,0.010964444114102257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,128,0.0032302221904198327
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,2048,0.00777511133088006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,512,0.019155555301242404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,16384,0.03960977660285102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,128,0.0041662222809261745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,2560,0.01035377797153261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,3072,0.02216000027126736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,32,0.0026497778793176017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,2560,0.006119110931952794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,128,0.008195555872387355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,128,0.004056000047259861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,32,7168,0.012074666718641916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,3072,0.11738755967881943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,5120,0.009979555176364051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,8192,0.03711466656790839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,7168,0.011197333534558615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,7168,0.031176000833511353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,128,0.0037768888804647657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,2048,0.006659555352396435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,512,0.007670222057236566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,4096,0.011377777490350934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,4096,0.008005333443482717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,32,0.003112000102798144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,16384,0.21406222714318168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,1024,0.0058168888919883305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,7168,0.010787555740939246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,16384,0.2598151101006402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,32,0.00619555554456181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,5120,0.01201777739657296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,4096,0.021817778547604878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,65536,0.25150932206047905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,768,0.0034168888297345904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,64,0.00701155596309238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,10240,0.022786666949590046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,3584,0.009487110707494948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,1536,0.5653893152872721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,768,0.007947555846638149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,3584,0.010295111272070143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,12288,0.04804177747832405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,32,0.004504000147183736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,256,0.005251555393139522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,3072,0.013167111410035027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,2048,0.006761777732107375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,1024,0.02896089024013943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,5120,0.00944800012641483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,3072,0.020641777250501845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,5120,0.006677333265542984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,3072,0.00869777798652649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,64,0.004878222114510006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,3072,0.006294222341643439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,2048,0.08693510956234402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,768,0.00490044429898262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,64,0.0028106667515304354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,8192,0.10646044545703465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,1024,0.004853333450025982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,4096,0.0075599998235702515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,5120,0.06659822331534491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,1536,0.004576888763242298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,3072,0.08036088943481445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,1536,0.006800888727108638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,6144,0.009649777577983009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,12288,0.02032355633046892
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,64,0.0025804444319672054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,256,0.003272888975010978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,768,0.007642666498819987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,768,0.009517333573765224
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,5120,65536,0.2782213422987196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,5120,6144,0.025460445218616064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,1536,0.00666133314371109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,128,0.011690666278203329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,32,0.0028595555987623003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,3072,0.006785777707894643
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,5120,0.2491075462765164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,128,0.007824888659848107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2048,8192,0.01716088917520311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,768,0.006045333213276333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,4096,0.00927111092540953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,16384,0.007354666789372762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,5120,0.288698673248291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,10240,0.01612711118327247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,5120,0.009396444592210982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,2560,0.013941332697868347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,1024,0.008406221866607666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,8192,0.015674667225943673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,8192,0.02296000056796604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,65536,0.1338311168882582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,3072,0.008858666651778752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,8192,0.0155404441886478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,1024,0.003896000070704354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,512,0.004549333204825719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,65536,10240,0.2020497719446818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,3072,0.11020266347461277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,512,0.010423111418883005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,32,0.005551110953092575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,512,0.00775911079512702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,12288,0.01538755496342977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,65536,0.19299199846055773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,32,0.002921777880854077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,3072,0.008542222281297049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,65536,2.4385503133138022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,12288,0.09161600139405991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,16384,0.0269813338915507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,2048,0.011053333679835001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,3584,0.023013333479563396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,12288,12288,0.21726755301157633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,16384,0.024688889582951862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,1024,0.00776533368561003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,2560,0.006758222149478064
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,5120,0.009215999808576372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,8192,0.05193066596984863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,6144,12288,0.03970222340689765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,2560,0.09034933646519978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,8192,0.012895111408498553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,512,0.005140444470776452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,16384,0.22088800536261666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,1024,0.009105777574910058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,7168,0.01942311061753167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,8192,0.02660266723897722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,7168,2048,0.010114666488435533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,128,2560,0.012299555871221753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,7168,65536,0.49530665079752606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,256,0.007713778151406183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,65536,0.12969688574473062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,4096,0.007354666789372762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,65536,0.08037955231136747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,128,0.004770666774776247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,4096,0.007299555672539606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,10240,0.8766986529032389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,512,0.008274666965007782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,10240,0.011662222445011139
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,1024,0.25991643799675834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,6144,0.023927999867333308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,32,0.005050666630268097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,7168,0.01701244380739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,32,0.007466666400432587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3072,2048,0.02938133478164673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,32,0.005462222215202119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,16384,0.024501333634058636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,32,0.020469332734743755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,10240,1536,0.01167377746767468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,2560,0.00795377790927887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,2048,0.007726222276687622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,2048,0.006531555619504716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,128,0.002704000100493431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,32,0.006158222340875202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,7168,0.22562665409511992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,10240,0.17956444952223036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,1024,0.005852444304360284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,12288,0.22897778617011177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,8192,0.015427556302812366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,3072,0.006689777804745569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,2048,0.23668710390726724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,768,0.0037973332736227247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,512,7168,0.026201777988009985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,1024,0.010320888625250923
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,2048,0.019039111004935372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,3072,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,768,16384,0.03811466693878174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,6144,0.03210399879349603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,32,0.012536889149083031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,6144,0.03921066721280416
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,768,0.03934755590226915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,3072,0.015798222687509324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,6144,1536,0.00814400033818351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,10240,0.05197777681880527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,256,4096,0.008443555898136562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,12288,0.054336888922585376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,3584,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,10240,0.36508978737725156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,2048,0.008480889101823172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,4096,0.008189332981904348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,65536,0.3086755540635851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,128,0.005062222066852781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,65536,8192,0.2535128858354357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,65536,0.08146222432454427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,768,0.009188444250159794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,1024,0.017295999659432303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,3584,0.03879822293917338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,512,0.007373332977294922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,128,0.0048942222363419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,65536,0.1150124470392863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,256,0.0035297779573334586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,10240,0.04068888889418708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,65536,128,0.020468443632125854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,128,3072,0.007335111498832703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,2048,0.011023999916182624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,128,0.006077333456940121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,512,0.013046222428480784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,65536,0.15392088890075684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,7168,65536,11.495024787055122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,256,0.0062773335311147906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,2048,0.012134222520722283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,64,0.0064311110311084324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,16384,0.040031111902660794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3584,64,0.006430222342411677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,65536,0.2271813286675347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,2560,0.02330044408639272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,128,0.005961777849329843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,2048,0.007037333316273159
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,1536,0.006486222147941589
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,6144,0.01461155547036065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,10240,0.027354665928416785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,768,5120,0.009335999687512716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,2560,0.036280890305837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,512,0.12292533450656468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,128,0.0042204442951414315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,16384,0.5054613219367133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,12288,0.05898400147755941
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,256,0.005774222314357758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1024,256,0.003168888803985384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,10240,0.009398221969604492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,512,0.00442755553457472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1536,8192,0.06961244344711304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,128,0.023960888385772705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,3072,0.0068711112770769335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,64,0.0026542221506436667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,1024,0.025263110796610515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,6144,0.010040000081062317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,32,0.028351111544503108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,6144,0.029620445436901514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,2048,0.0159582214223014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,5120,0.013800889253616333
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,3072,0.012607110871209038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,16384,0.025250666671329077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,512,0.006125333408514659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,1536,0.015191111299726697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,12288,0.030128889613681372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,5120,0.00683111117945777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,10240,4096,0.022448889083332483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,2048,0.006514666808976068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,6144,64,0.008242666721343994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,7168,0.01533422205183241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,3584,0.009858667022652095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,2560,0.012548444171746572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,256,0.004679111142953237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,16384,0.015408888459205627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,10240,10240,0.05340800020429823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,256,0.029729776912265356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,256,0.0029537777105967202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,6144,0.8014231257968478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,768,0.006708444406588872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,2048,0.016163556112183463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,1536,0.01456622282663981
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,1024,0.00573244442542394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,1024,0.030537777476840552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,3584,0.07470044162538317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,64,0.002571555475393931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,10240,0.03717155588997735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,3072,0.04157333241568671
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,3072,0.07236088646782769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,16384,0.028981334633297388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,2560,0.006400888992680444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,8192,3584,0.01910666624704997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,256,0.0031759999692440033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,3072,0.02558311157756382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,512,0.004381333374314838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,256,0.002992000016901228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,128,0.005843555761708154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,768,0.007281777759393056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,512,0.0032417778339650896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,256,0.004621333132187526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,2048,0.007101332975758447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,6144,0.00739644467830658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,10240,0.12952266799079046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,7168,0.01094133324093289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,2048,0.013456888496875763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,256,768,0.006007110906971826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,64,4096,0.00573511090543535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,8192,0.012230222423871359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,3072,0.021839110387696162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,2048,0.005342222336265776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,2560,0.01124888906876246
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,32,0.004921777794758479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,256,0.003953777667548922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,3072,0.007425777614116669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,2048,0.006275555739800136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,128,0.0029768887907266617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,65536,32,0.031116445859273274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,256,0.005375111268626318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,65536,0.41444889704386395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,768,0.007872888611422645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,32,0.0037626665499475268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,768,0.005912000106440649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,16384,0.01919644408755832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,16384,0.023992000354660883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,8192,0.014780445231331719
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,65536,0.09176711241404216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,2560,0.007080000307824876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,5120,0.018557333283954196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,4096,0.0083155557513237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1536,8192,0.03377066718207465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,7168,0.01074577785200543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,7168,0.019627556204795837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3584,3584,0.011792889071835412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,7168,0.04387466775046455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,5120,0.015609777635998197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,8192,0.007366221812036302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,64,0.00434666665063964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,16384,0.05890311135186089
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,2048,0.00592533333433999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,5120,7168,0.016369778248998854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,4096,0.024023110667864483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,64,0.003478222423129611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,5120,0.01071999967098236
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,8192,0.539355542924669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,128,0.02974933385848999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,4096,0.028424888849258423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,3584,0.012270222107569376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,16384,1.1954124238755968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,5120,0.012845333251688214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,10240,0.03522133496072557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,1024,0.003698666476541095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,65536,0.10770844750934178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,1536,0.011807111402352652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,32,0.005036444299750858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,1024,0.005769777629110549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,1024,0.0052382221652401825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,4096,0.007721777591440413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,3072,6144,0.04092622134420607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,8192,0.016878222425778706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,2560,0.009639999932712978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,6144,0.20411822530958387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,65536,6144,0.429528898662991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,4096,0.01910044418440925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,256,0.005856888989607493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,1536,0.008956444760163626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,16384,0.036545776658587985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,768,0.005707555760939916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,5120,0.019166222876972623
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,5120,0.2133457793129815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,512,0.005443555613358815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,768,0.007921777665615082
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,1536,0.006698666761318843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,65536,8192,0.14966489209069148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,7168,0.014836443795098199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,4096,0.008453333543406593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,5120,0.009633777870072259
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,32,0.002850666642189026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,32,0.03562044435077243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,8192,0.01704355577627818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,12288,1024,0.012033777932325998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,256,0.004208000169859992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,256,0.0038293335172865125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,5120,0.009667555491129557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,4096,0.034557332595189415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,2048,0.007807110746701558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,6144,0.012897777888509961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,128,12288,0.007849777738253275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,16384,0.014653333359294467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,12288,0.018607111440764535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,32,0.0028400001012616684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,256,0.003516444315512975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2048,16384,0.02234577801492479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,3072,0.013259555730554791
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,8192,12288,0.04056977894571092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,2048,0.006788444601827198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,16384,0.020871111088328894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,1024,0.0041804446114434134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,3584,0.11170400513543023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,12288,0.013348444468445249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,16384,2048,0.015930665863884818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,768,0.003719999972316954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,2048,0.007710222568776872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,8192,0.015003555350833468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,3584,0.00656088896923595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,1536,0.012809777425395118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,65536,2048,0.7324435445997449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,16384,0.8627289136250814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,32,0.0058000000814596815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,4096,0.022224000758594934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,5120,0.014925332532988654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,7168,0.019853333632151287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,512,0.005677333308590784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,6144,0.044953776730431445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,2560,0.008534222013420528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1536,3072,0.009422221945391761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,768,12288,0.011319999893506369
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,256,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,2560,0.006833777659469181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,4096,5120,0.01679111189312405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,6144,0.07021688752704196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,10240,0.07359911335839166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,128,0.013156444662147097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,65536,0.21684711509280732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,4096,0.008244444098737504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,3072,0.005902222047249476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,3072,0.014821334017647637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,512,0.005616000129116907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,5120,0.006897777732875612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,1536,0.008034666379292807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,6144,0.019541333119074505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,6144,0.05126488871044583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,512,0.007879999776681265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,12288,0.014511111709806653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,256,0.0047146665553251905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,32,0.006301333506902059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,768,0.006405333264006509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,16384,3584,0.030448890394634668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,8192,0.2754017776913113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,64,0.005868444426192178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,3584,0.012461333639091916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,1536,0.007095110913117726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,5120,0.012059555285506778
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,1536,0.006181333214044571
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,10240,0.015013333823945789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,65536,6.639581468370225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,128,0.007848888635635376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,12288,0.0662026670244005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,65536,0.2992275555928548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,768,0.00795199970404307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,8192,4096,0.42996711201137966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,128,0.003343111111058129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,768,0.005118222286303838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,65536,0.8403724564446343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,128,0.006132444573773279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,2048,0.015990222493807476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,512,0.004540444248252445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,256,0.030382222599453394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,256,0.03272622161441379
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,6144,1536,0.009167110754383935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,3072,0.007253333098358578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,65536,0.08655199739668105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,1024,0.008991111483838823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,768,0.004373333520359463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,6144,0.014523555835088095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,4096,65536,0.097126219007704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,768,0.02612000041537815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2048,3072,0.10163822438981797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,64,0.006212444355090459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,10240,0.07570755481719971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,1024,0.008797333472304875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,256,0.006514666808976068
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,7168,0.013939556148317126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,3584,0.015378667248619927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,6144,0.05981777773963081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,5120,0.2374702294667562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,4096,0.02310933338271247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,7168,0.010517333116796283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,6144,0.01388177772363027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,3072,0.033260444800059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,65536,0.14652800559997559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,16384,0.07740088966157702
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,5120,0.022484444909625586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,12288,0.06670400169160631
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,7168,0.018248889181349013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,128,0.021546666820844013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,4096,0.04342577854792277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,12288,0.015967110792795818
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,4096,128,0.004827555682924059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,6144,0.009482666850090027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,4096,0.040305776728524104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,256,0.012925333446926542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,1024,0.007129777636792924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,64,0.004139555411206352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,256,0.004846222284767362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,4096,0.02308622168170081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,4096,6144,0.01759288873937395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,4096,0.010290666586822933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,2560,0.0070977773931291364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,6144,0.0464826656712426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,3584,0.009889778163697984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,2048,0.021800888909233943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,4096,64,0.006551110910044775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,2048,0.01033866653839747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,6144,0.026512889398468867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,8192,0.03285955720477634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,64,0.005249777601824866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,8192,0.01808177762561374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,1024,0.006088888893524806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,2048,0.010675555302037133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,8192,0.01347733371787601
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,3072,0.007697777615653143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,2560,0.008068444828192392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,3072,0.00941244430012173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,128,0.009946666657924652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,16384,0.041209777196248375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,128,0.006877333339717653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,2560,0.007272000114123027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,3072,0.008325333396593729
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,5120,0.012168889244397482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,12288,0.04233688778347439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,128,0.0026071110947264563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,2560,0.007175111108356052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,32,0.00499911109606425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,64,0.002432000098956956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,4096,0.0281644430425432
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,3584,0.008857777549160851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,1024,0.005234666582610872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,6144,0.03527199890878465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,6144,0.07007377677493624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,16384,0.01902755598227183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3072,3072,0.010090666512648264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,8192,0.022853333089086745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,3584,0.025106666816605464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,7168,3584,0.012832000023788877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,7168,0.02051377793153127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,256,0.003039111072818438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,64,0.006017777654859755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,2560,0.007246221933099959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,12288,0.6450906859503852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,1536,0.0063786668082078295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,7168,0.135587559805976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,256,0.04941244588957893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,8192,0.006627555522653792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,128,512,0.00452977791428566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,768,0.005612444546487596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,3584,0.013677332964208392
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,7168,0.013424889081054263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,6144,0.014114666316244336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,2560,0.08933333555857341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,65536,0.10584266980489095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,65536,0.06927111413743761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3584,6144,0.04648800028695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,65536,4.079656812879774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,3072,0.007310222420427535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,768,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,2560,0.037485334608289934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,65536,0.23276978068881563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,8192,256,0.003346666693687439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,3072,0.010779555473062726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,2560,0.008316444853941599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,6144,0.06296622090869479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,10240,0.048102223210864596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,3072,0.06467466884189181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,768,64,0.002624888800912433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,10240,256,0.008577778107590145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,768,0.07863466607199775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,6144,0.019648889700571697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,3072,0.007625777688291337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,32,0.004132444245947732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,512,0.003306666596068276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,10240,0.015499555402331881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,768,0.006083555519580841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,768,0.007002666592597961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,10240,0.06569244464238484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,512,0.024717332588301763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,64,0.03807377815246582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1024,12288,0.016530666086408827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,768,0.02087200019094679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3584,3584,0.08738222387101914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,4096,0.007659555309348636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,4096,256,0.004751111070315043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,5120,0.03770844472779168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,3584,0.01776977711253696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,256,0.005406222409672207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,3584,0.007880888879299164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,2048,0.006812444577614467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,65536,0.4284675651126438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,10240,0.08169244395362006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,5120,0.012225777738624148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,10240,0.013481777575280933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,5120,0.017843556072976854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,10240,0.01369866645998425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,256,0.0033297776761982176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,7168,0.006124444305896759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,1024,0.008026666939258575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,5120,0.05171289046605428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,3072,0.009610666996902889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,64,0.006089777582221561
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,1024,0.004051555362012652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,32,7168,0.005879111174080107
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,10240,0.006700444552633498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1536,8192,0.014616888430383472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,128,0.004979555391603046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,4096,0.0057848890622456866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,64,0.0047457776963710785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,1536,0.005983110931184556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,65536,0.23723999659220377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,1024,0.04405422343148125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,65536,0.6843697759840223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,3072,0.01566488875283135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,768,0.012764444781674279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,512,0.0046746668716271715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,1536,0.007729777859316931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,7168,0.01184177812602785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,16384,0.019089778264363606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,16384,0.04086666636996799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,10240,0.22193066279093424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,10240,0.08255022101932101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2048,16384,0.02292799949645996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,1536,0.009693333672152625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,5120,0.04614666766590542
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,8192,0.018223999275101554
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,16384,0.0351235568523407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3584,16384,0.03613866700066461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,7168,0.017660443981488545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,1536,0.026447110705905493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,5120,0.00673333348499404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3072,7168,0.013750221994188098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,8192,0.030011554559071858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,512,3072,0.006253333141406377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,1024,0.012235555383894177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,256,0.005420444326268301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,1024,0.006451555424266391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,6144,1024,0.006353777729802662
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,128,0.0028311111446883944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1024,2048,0.02700977855258518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,768,6144,0.009581333233250512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,10240,1024,0.012844444149070315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,8192,0.011491555306646558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,12288,0.032765332195493914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,65536,3584,0.11070133580101861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,2560,0.05576088693406847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,2560,0.006728888799746831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,5120,0.008607111043400234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,3584,0.018699554933442008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,4096,0.006214222146405115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,768,0.0046897778908411665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,32,0.004581333448489507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,64,0.007172444628344641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,64,0.0047679998808436925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,4096,0.008468444148699442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,64,0.0027386666172080566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,1024,0.005105777747101254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,3584,0.017159110969967313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,1536,0.010119111173682744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,10240,0.009732444253232744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,6144,0.012201777762836881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,128,0.0027919999427265594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2048,1024,0.0063324446479479475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,6144,0.20419111516740587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,10240,0.007324444750944774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,1536,0.00979733301533593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,128,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,256,0.003936000168323517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,1536,0.010659555594126383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,32,0.011746666497654386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,4096,0.08295377757814196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,10240,0.02913155489497715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,3584,0.018144889010323417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,4096,0.016599999533759225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,6144,0.02247911029391819
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,2560,0.00980088859796524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,8192,0.013787555197874704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,32,768,0.005745777653323279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,65536,0.27954665819803876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,4096,0.02049422264099121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,1024,0.2450631194644504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,128,0.00598933340774642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,64,0.002944888960984018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,512,5120,0.00926755534278022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,12288,0.017422222428851657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,64,5120,0.010545777777830759
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,10240,0.18845866786109078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,1024,0.005997333261701796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,768,0.015523556205961438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,1024,0.005836444596449534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,4096,0.020836444364653692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,2048,0.008443555898136562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,256,0.005745777653323279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,65536,0.34783466657002765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,256,0.006253333141406377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,256,0.00943199959066179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,64,0.004977777600288391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,1536,0.010749333434634738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,4096,0.0085235552655326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,1024,0.00566133318675889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,10240,0.07765866650475396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,768,0.004716444346639845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,8192,0.013456000222100152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,256,0.00464088883664873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,1536,0.004323555363549126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,4096,0.009830222361617619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,7168,0.016350222958458793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,32,0.0027520000520679685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,7168,0.006462222172154321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,65536,0.02970577942000495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,3072,0.011389333340856763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,3072,0.06813510921266344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,8192,0.008827555510732863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,8192,0.04077333211898804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,16384,0.03371999992264642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,7168,16384,0.03952266772588094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,3584,0.021176889538764954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,2560,0.013649777405791812
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,12288,3584,0.47692357169257266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,512,2048,0.007003555695215861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,64,0.0033520000676314035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1024,128,0.0042160000238153665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,1536,0.010648888846238455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,6144,0.025077333052953083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,32,0.0025439999169773525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,5120,0.06795110967424181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,64,0.002583111118939188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,16384,0.12563021977742514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,2560,0.006317333214812809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,3072,0.0105608892109659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,2560,0.008311111066076491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,4096,0.15648889541625977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,2048,0.009932444327407414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,32,0.003105777833196852
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,2560,0.006752000086837345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,3584,0.010821333361996545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,1536,0.010612444745169746
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,768,0.0035626664757728577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,768,0.005420444326268301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,512,0.038683556848102145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,3584,0.019110222657521565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,768,0.036537776390711464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,65536,0.7877288924323188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,1024,0.008084444536103142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1024,4096,0.011373333632946014
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,65536,0.2089093393749661
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,128,0.002982222164670626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,12288,6144,0.026203556193245783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,12288,0.04404533240530226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,512,0.005685333162546158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,3584,0.018425777554512024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2048,1536,0.0058515556156635284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,1536,0.006404444575309753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,2560,10240,0.020282665888468426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,10240,0.023056889573733013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,7168,128,0.011350221931934357
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,1024,0.0052826665341854095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,3072,0.020820443828900654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,1024,0.0069120000633928515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,7168,4096,0.1842648850546943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,4096,256,0.0037226668662495087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,16384,0.3542986710866292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,32,0.0030657777355776895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,65536,0.8657253583272299
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,7168,0.030978666411505804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,4096,0.01278222186697854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,10240,0.03229422370592753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,1024,0.005558222118351195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,256,3584,0.048862222168180675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,32,0.009007111191749573
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,64,0.007180444068378872
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,8192,0.6266329023573134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,2560,0.03516355488035414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,7168,0.4681084420945909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,6144,1024,0.013463999662134381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,5120,0.013983999689420065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,6144,0.08641066816118027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,32,0.003992889076471329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,128,512,0.004563555535342959
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,2048,0.03752711084153917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,3072,0.020459555917316012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,128,768,0.013912889692518445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1024,10240,0.014492443866199918
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,10240,0.014101333088344999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,5120,1024,0.005667555663320754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,1024,0.008798222574922774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,64,0.003039111072818438
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,256,0.004054222255945206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,3072,0.007330666813585494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,768,0.004062222109900581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,3072,0.00742222203148736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,3072,0.18765421708424887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,1024,0.012907555533779992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,65536,0.07535555627610949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,5120,1024,0.005782222168313132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,768,16384,0.07606399721569486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,5120,0.00852177788813909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,3584,0.011226666470368704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,2048,0.014142221874660917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,6144,0.045443554719289146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,512,0.005336888962321811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,4096,128,0.006308444258239534
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,768,0.05655022462209066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1024,6144,0.023518222901556227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,2048,0.007171555525726742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,2048,0.01090400003724628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,7168,0.007167110840479533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,12288,0.7615217632717557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,1536,0.008014221986134848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,3072,0.013959111438857185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,32,0.024357333779335022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,12288,0.02079288827048408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,3584,0.011202666494581433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,3584,4096,0.012887111140622033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,1024,0.00665244460105896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,5120,0.00723555518521203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,16384,0.09964711136288112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3072,512,0.004120888809363048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,512,32,0.005535999933878581
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,12288,0.046592000457975596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,4096,0.009471110999584198
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,5120,0.0118568887313207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,64,3072,0.015793777174419828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,8192,0.26275377803378636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,1536,0.014829332629839579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,3584,0.011957333319716983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,2048,0.014448000325096978
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,16384,0.06077510780758328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1024,10240,0.02090666691462199
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,65536,0.17784889539082846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1024,8192,0.014766222900814481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,512,10240,0.03916710946294997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,1536,0.006113777971929974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,65536,0.19540799988640678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,65536,0.1962906652026706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,2048,0.006778666542636023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,7168,0.009296000003814697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,256,0.03844088978237576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,4096,0.023409777217441138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,1536,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,7168,0.02608888844648997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,128,64,0.002842666581273079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,3584,0.0610426664352417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,1536,0.016269332832760282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,5120,0.014398222168286642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,8192,0.011063111325105032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,64,0.008759111166000366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,10240,0.06659199794133504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,5120,0.38580621613396543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,3072,0.07839733362197876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,512,0.003863999827040566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,1024,0.015814221567577787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,1536,0.01109866632355584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,10240,0.05848089191648695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2048,1536,0.005658666706747479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,512,0.006574222197135289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,3584,0.006588444527652528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,2048,0.06451555755403307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,6144,0.010152888794740042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,7168,0.017490666773584154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,64,2560,0.014888889259762235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,768,2560,0.007616000043021307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,7168,0.3027404414282905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,64,0.03975111246109009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,12288,256,0.10047466887368096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,5120,0.015856888559129503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,4096,0.006717333363162146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,12288,0.029639999071757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1536,10240,0.018453333112928603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,2048,0.011571555501884885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,4096,16384,0.033165332343843244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3584,128,0.003048000029391713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2560,256,0.003488000068399641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,6144,2.604722764756944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2048,32,0.004828444371620814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,1536,0.007924444145626491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,3072,0.010175111393133799
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,3072,0.008314666648705801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,2048,0.01387288918097814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,128,16384,0.006225777582989798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,4096,2048,0.009230222139093611
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,768,10240,0.19149777624342176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,65536,0.1935288906097412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,3072,0.0070159998204973005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,8192,0.24247201283772787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,8192,8192,0.024745777249336243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,1024,0.013360888593726687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,32,0.0026817777090602448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,2048,0.030951112508773804
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,3584,0.012714666624863943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,768,0.005378666851255629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,64,0.0053377776510185665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,7168,0.04079644547568427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,6144,0.16579911443922254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,5120,0.11058755715688069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,128,768,0.005828444328573015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,128,0.005201777650250329
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,1536,0.009873777627944946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,1536,0.036928001377317644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,3584,0.0103502223889033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,256,512,0.0033031110134389666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,2560,0.007589333587222629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,8192,256,0.00942133367061615
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,7168,0.036215109957589045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,256,0.006505777852402792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,3584,0.09460088941786025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,6144,0.0096124443742964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,768,0.009581333233250512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,5120,0.009737778041097853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,65536,0.6152826415167915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,768,0.005396444350481033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1024,7168,0.010497777826256223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,5120,0.0063457778758472866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,5120,0.021506667137145996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,4096,0.019293333093325298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,5120,0.009543110926946005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,1536,0.01591377788119846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,768,0.043566223647859365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,6144,0.04953866534762912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,64,10240,0.013750221994188098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,64,0.003085333233078321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,128,0.004671999977694618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,12288,0.03947644432385763
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,4096,0.01161688897344801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,12288,0.0163137780295478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,65536,0.4796097543504503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,256,0.0041368889311949415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,1024,0.006433777925040986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,2048,0.024414221445719402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,16384,0.04186133212513394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,1536,0.014764444695578681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,2560,0.019274666905403137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,12288,6144,0.040390223264694214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,768,0.012272000312805176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,2048,0.005959110955397288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3072,512,0.004095111042261124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,512,0.005579555614127054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,768,16384,0.023358222511079576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,1536,0.015623999966515435
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,5120,0.1578622261683146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,256,512,0.003294222263826264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,512,0.005037333402368757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,3584,0.008600888980759515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,10240,10240,0.6468426916334364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,2560,0.010716444088353051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,8192,0.06425778071085612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,1536,0.02342844506104787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,4096,0.010860444770918952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,6144,0.01868355605337355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,256,0.0074373334646224976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,512,0.007055111229419708
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,2048,0.04431110951635572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,64,1024,0.003948444293604957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,64,0.005326222214433882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,3584,0.03868622250027127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,512,0.005009777843952179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,6144,0.011420444481902652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,1536,0.006108444597986009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,64,0.002833777831660377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,2048,0.006693333387374878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,4096,0.018165333403481376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,8192,0.0062773335311147906
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,1024,0.015231110983424716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,128,0.3799440066019694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,65536,256,0.11334578196207683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,32,0.011734222372372946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,64,0.007127111156781514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,32,2048,0.01094044413831499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,2560,0.03812266720665826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,256,6144,0.01072533345884747
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,1536,0.06232711341645983
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,3584,0.012640888492266336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,2560,0.010272888673676385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,6144,0.03267822331852383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,256,0.005769777629110549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,6144,0.04064888755480448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,3584,0.011678222152921887
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,64,0.002761777697337998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,128,0.004964444372389051
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,2048,0.006640000061856375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,3072,0.018248889181349013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,2048,0.03527288966708713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,2560,0.006337777607970768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,6144,0.027073777384228174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,1024,0.00527555536892679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,5120,0.04093155595991347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,5120,1024,0.018357333209779527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,256,0.005719999886221356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,12288,0.24112534523010254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,12288,0.025405334101782903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,1024,0.012092444631788464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,128,0.01664533383316464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,128,0.005071111023426056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,8192,0.03365866674317254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,6144,0.04452355702718099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,5120,6144,0.020440000626775954
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,10240,0.01663555536005232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,5120,0.013450666434235044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3584,3584,0.0124177775449223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,16384,0.09087910917070176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,7168,0.01833777791923947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,768,65536,0.23000976774427626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,64,0.005193777796294955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,10240,0.027503111296229895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,65536,0.3112257851494683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,10240,0.5670284695095485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,8192,0.03549244337611728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,3072,0.007626666790909237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,10240,0.031614220804638334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,12288,2048,0.011569778124491373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,2560,0.016019556257459853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,5120,0.017879999346203275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,4096,0.02092444399992625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,128,0.0038364442686239877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,10240,0.015973332855436537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,10240,0.11107289128833348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,3072,0.008479110896587372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,12288,0.10400711165534125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,10240,0.017096888687875535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,10240,0.040736890501446195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,1024,0.005164444446563721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,256,0.021264889174037512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,128,0.009696000152164036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,7168,0.006039111150635614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,7168,0.012048000262843238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,5120,0.009054222040706212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,8192,256,0.006454222318198945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,1024,0.006510222123728857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,5120,0.010503999888896942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,2560,0.009373333719041612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,1536,0.0059004442559348206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,7168,512,0.013279111021094851
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,5120,0.024499555428822834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,6144,0.009617778162161509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,3584,0.010001777774757808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,32,0.008466666771305932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1024,7168,0.008942222429646386
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,10240,0.03289422061708238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,768,0.3386497762468126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2560,4096,0.02293777796957228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,2560,0.022611555125978258
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,512,0.006551110910044775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,8192,1024,0.037244445747799344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,256,0.003028444531891081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,2560,0.03188177943229675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,32,0.008737777670224508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,768,0.007825777762466008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,64,0.012636444634861417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,32,2048,0.012905778156386482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,7168,12288,0.029118223322762385
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,2048,0.007220444579919179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,1536,0.004590222405062782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,2048,0.0058808889653947614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,12288,0.030864887767367895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,32,7168,0.008395555946562026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,2048,0.012309333516491784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,12288,8192,0.05390222205056084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,65536,0.01517955627706316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,768,0.14991733762953016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,8192,0.012245333029164208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,256,0.005571555760171678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,1536,0.012011555333932241
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,6144,0.02535022298494975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3584,256,0.02680177821053399
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,10240,0.028115553988350764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,10240,0.03776177763938904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,65536,768,0.02181511123975118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,2048,0.7362426651848687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,64,0.006707555717892117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,768,10240,0.014745778507656522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,7168,0.16932266288333467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,256,0.003628444340493944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,6144,0.006792888873153263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,8192,10240,0.030821333328882854
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3584,512,0.005203555441564984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,256,0.007882666256692674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,6144,0.019444444113307528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,10240,0.03356711069742838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,128,0.016885333591037326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,512,0.005586666779385672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,6144,0.025377778543366328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,32,0.002550222186578645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,16384,2560,0.021197333931922913
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,128,128,0.003960000144110786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,16384,0.6172809071011013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,1536,0.007688889073001013
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,1024,0.004480000171396467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,65536,0.15122400389777288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,1024,0.005979555762476391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,5120,7168,0.027806222438812256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,3072,7168,0.014648000399271647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,16384,0.010296000374688042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,768,0.00573244442542394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,65536,0.10984444618225098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,1024,0.005035555611054103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,2560,0.007807999849319458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,32,0.00498488876554701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,4096,0.016562667157914903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,7168,0.01092711091041565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,6144,0.02984444300333659
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,8192,0.07221066951751709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2048,12288,0.01833511061138577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,768,0.013136000268989138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,2048,0.009153777526484596
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,65536,256,0.05312711000442505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,128,1024,0.005912000106440649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2560,64,0.00408000002304713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,5120,0.03609422180387709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,2560,0.013112000293201871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,256,1024,0.009199110998047723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2048,8192,0.01863466699918111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,3584,0.012186666329701742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,512,6144,0.009889778163697984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,10240,0.05619555711746216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,8192,2560,0.011297778122954898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,7168,1024,0.009932444327407414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,768,0.01018133345577452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,2560,0.04773422082265218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,1024,0.006110222389300664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,3584,0.06610488891601562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,6144,1.2274977366129558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,64,0.006554666492674086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3072,65536,0.08760799964269002
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,4096,1.6852471033732097
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,7168,0.011604444848166572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,6144,0.012753778033786349
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,256,4096,0.01735733283890618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,10240,0.015024888846609326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,12288,0.044826666514078774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3584,5120,0.013343999783198038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,1536,0.08401422368155585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,256,7168,0.010535111029942831
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2560,256,0.005572444448868434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,3584,0.010195555786291758
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,3584,0.011448000040319232
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,8192,512,0.004865777575307422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,5120,128,0.01388088862101237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,65536,0.10083288616604275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,1024,0.028033776415718928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,3584,0.01182577759027481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,2048,0.009903999666372934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,7168,0.049026668071746826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,256,0.004612444589535396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,128,0.005602666487296422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,128,0.004335111214054955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,4096,0.03535288903448317
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,5120,0.010168000227875179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,64,0.04923289020856222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,1536,7168,0.016151999433835346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,65536,512,0.023356444305843774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,768,7168,0.008169777691364288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,65536,0.04409511221779717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,4096,0.013599110974205865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,65536,0.20892800225151908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,4096,0.020968000094095867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,2048,0.012914666699038612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,8192,0.713199085659451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,16384,0.10672266615761651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,12288,0.8751813570658366
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,768,0.007787555456161499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,12288,3584,0.28634312417772084
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,16384,0.035409778356552124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,128,4096,0.03548533386654324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,5120,0.023638221952650283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,6144,0.01349422252840466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,2048,0.006146666490369373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,8192,3072,0.016597333881590102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,2560,0.0064195555945237475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,12288,0.018744000130229525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,16384,2560,0.016407110624843173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,65536,0.4975004196166992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,4096,3584,0.0414319998688168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,12288,0.04667377803060743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,768,0.007476444045702617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,32,0.00332266671790017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,65536,0.30600354406568736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,3072,0.08982044458389282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,1536,0.0069368887278768755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1536,8192,0.01607911123169793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,6144,0.023711111810472276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,768,0.008832889298597971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2560,256,0.006928000185224745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,32,0.0026622222115596137
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,4096,8192,0.04772355490260654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,8192,0.864467567867703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,10240,0.27158045768737793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,16384,0.00831733312871721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,16384,0.027760000692473516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,256,0.0032257777121331957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,8192,10240,2.0542818705240884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,12288,0.057998220125834145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,256,0.004912000149488449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,128,256,0.0047848886913723415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,3072,0.007063111497296228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,12288,1.1381147172715929
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,64,0.0029893333299292457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,512,0.005409777578380372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,3584,128,0.004278222305907143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,64,0.010151111417346531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,5120,0.021895999709765118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,12288,3072,0.026855111122131348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,65536,0.6096320152282715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,8192,0.7652639812893338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,2560,65536,0.08896622392866348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,7168,0.03446311089727614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,5120,0.009740444521109263
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,65536,512,0.02991822361946106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,256,2560,0.006123555617200003
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,65536,0.06947466399934557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,8192,0.037335998482174344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,16384,0.02366666661368476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,3584,0.008660444782839881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,128,0.008486222061845992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,10240,8192,0.44562753041585285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,512,3072,0.0063440000845326324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,3584,0.14280800024668375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2560,16384,0.06291111310323079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,6144,1024,0.006508444332414203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,12288,6144,0.14987644884321424
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,32,0.017055999901559617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,5120,0.008151111503442129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,4096,0.007117333511511485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,32,32,0.0025822222232818604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,768,0.0034933334423436057
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,3584,0.02036977807680766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,512,0.008069333102968004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,256,0.005643555687533484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,7168,0.08559377988179524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,8192,0.2700800100962321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,1536,0.0064026667839950984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,16384,32,0.003944888710975647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,64,7168,0.011959111524952782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,32,0.0033031110134389666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,4096,0.13082488377888998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,1536,0.026736888620588515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,3072,0.03618844350179037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,65536,2.842277314927843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2560,256,0.0038737778862317405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,7168,0.02308888898955451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,128,0.005200888961553574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,16384,0.024113777610990737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,5120,2560,0.010312889185216691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,32,0.005108444641033809
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,65536,0.09765066703160603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,10240,0.03758488761054145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,6144,0.8185253143310547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,1536,0.013008000122176277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,4096,0.012524444195959302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,10240,4096,0.028669334120220605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,2048,0.00924711094962226
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,10240,0.032885332902272545
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,7168,0.2713395489586724
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,512,0.004302222281694412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,4096,0.05170489019817776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,6144,1536,0.007844444778230455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,256,0.004618666652176115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,64,0.003185777821474605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,10240,0.20949244499206543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,3072,0.007648888561460707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,3072,0.007615110940403408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,16384,0.019161777363883126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,8192,4096,0.014167111780908374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,768,0.005909333212508096
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,512,512,0.00443733317984475
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,64,0.006529777828190062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,32,6144,0.0069439998931354946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,10240,3072,0.015871110889646742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,768,0.004998222407367495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,768,0.09840799702538384
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,1024,2560,0.00996977753109402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,7168,0.009524444739023844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,7168,10240,0.042281776666641235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,3072,0.019719999697473314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,128,256,0.006614222294754452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,12288,0.018274666534529794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,7168,0.017475555340449016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,256,0.017703110973040264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,16384,0.08213688929875691
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,5120,0.011759999725553723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,2560,0.011760888828171624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1536,8192,0.01183111137813992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,32,0.007680888805124495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,7168,0.034550223085615374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,128,2560,0.006822222222884496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,8192,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,5120,0.17141689194573295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,16384,10240,0.046048889557520546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,2048,0.01568266749382019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,256,0.005342222336265776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,8192,0.02058577703105079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,3584,0.007471111085679796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,1024,0.019987555013762582
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,5120,0.009570666485362584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,65536,0.24847377671135795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,10240,0.013921777407328287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,12288,0.015962666935390897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,10240,0.021912000245518152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,64,0.008595555192894405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1024,6144,0.01166488892502255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,64,0.008672888908121321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,256,7168,0.007314666277832455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,1536,0.0064106666379504735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,7168,0.09806578026877509
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,4096,4096,0.01332266628742218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,3584,0.1115315490298801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,7168,0.17980178197224936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,3584,0.029723554849624634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,3072,0.010287111004193624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,768,0.01551644504070282
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,256,0.005709333138333426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,1024,0.09052089187833999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,16384,0.05479733480347527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,10240,0.0074346669846110875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,768,0.003683555457327101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,6144,0.025176000263955858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,1024,0.0087013335691558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,7168,6144,0.02936977810329861
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3072,3584,0.06201955344941881
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,768,0.005698666804366642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,512,0.005558222118351195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,64,0.005618666609128316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,2560,0.005965333431959152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,768,0.0036195553839206696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,3072,0.01627200014061398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,512,0.03801777627733018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,64,65536,0.013179555535316467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,3584,0.01293511109219657
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1536,128,0.003220444545149803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,64,16384,0.021535999245113794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,2048,0.006821333534187741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,65536,0.3526346683502197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,12288,0.040656887822681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,3072,8192,0.018755555152893066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,12288,0.022304889228608873
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,1536,0.01995022263791826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,3072,0.01440177857875824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,32,32,0.002576888849337896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,5120,0.013334222137928009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,3072,0.28039733568827313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,4096,0.007937777373525832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,512,10240,0.010474666953086853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,3584,0.1139262252383762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,5120,0.031133333841959637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,12288,0.02159822152720557
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,512,1024,0.004951111144489712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3072,65536,0.10106044345431858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,2560,0.00979644474056032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1024,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,10240,0.02139288849300808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,3584,0.021898667017618816
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,65536,0.06689244508743286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,256,0.00906133320596483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,768,65536,0.5580879847208658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,16384,2560,0.02253955602645874
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3072,3584,0.01462933255566491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,512,0.03123377760251363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,6144,0.019468444916937087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,1024,0.004139555411206352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,512,0.0053297777970631914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,64,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,5120,0.028592000404993694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,12288,1024,0.009108444054921469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,4096,0.01867999964290195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,3072,0.008762666748629676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,4096,0.009332444104883406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,4096,0.06962044371498956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,768,0.019713777634832595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,65536,16384,0.2763857841491699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,10240,0.07860533396402995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,12288,0.01919555498494042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,768,128,0.004067555483844546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,4096,0.013561777770519257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,3584,0.01606044504377577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,12288,0.08101689153247409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,8192,0.031161778502994116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,4096,0.008225777910815345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,6144,4096,0.013236444857385425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,64,5120,0.00960000024901496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,32,0.0031795555518733132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,65536,0.06755644745296903
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2560,3584,0.010234666367371878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,2560,0.010140444669458602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,7168,0.010844444235165914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,64,0.002773333340883255
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,256,0.008774222599135505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,5120,0.008176888856622908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,65536,4.491401672363281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,128,5120,0.012096000214417776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,2560,0.019898666275872123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,7168,0.016931555337376065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,2560,0.014912888407707214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,5120,256,0.026940445105234783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,10240,0.09230399794048733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,12288,2.5699520111083984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,6144,0.029284444120195176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,5120,0.013944000005722046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,3072,0.18775910801357695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,6144,0.1638560030195448
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,3072,0.014384888940387301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,768,1536,0.008397333323955536
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,256,0.006912888752089606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,4096,65536,6.474504682752822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,2560,0.011427555647161273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,1024,0.007625777688291337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,3584,0.019511999355422128
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,1024,0.008598222500748104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,768,512,0.005424889011515512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,4096,128,0.008636444807052612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,768,65536,0.10218133529027303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,768,0.005952888892756567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,16384,2560,0.9310355716281467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,7168,0.013285333083735572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3584,3584,0.012091555529170565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,64,5120,0.010410666465759277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,8192,0.05540711018774244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,4096,0.010880000061459012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,12288,0.025123554799291823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,3584,0.009550222092204625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,10240,0.013951111170980664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,65536,0.17859110567304823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,128,0.0075786668393347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,128,768,0.0046640001237392426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,1024,0.020940444535679285
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,128,0.00639111093348927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,128,0.013608888619475894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,512,2560,0.007763555480374231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,768,12288,0.01606399979856279
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,128,0.002700444517864121
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,6144,0.006072000082996156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,768,0.00565244464410676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2560,3072,0.1710453298356798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,6144,0.024289778537220422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,3072,0.008779555559158325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,5120,0.19769510957929823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,1536,0.024485333098305598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,5120,0.4621822039286296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,7168,64,0.009011555876996782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,4096,2048,0.031593776411480375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,1536,0.007409777906205919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,768,0.00847111145655314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2048,7168,0.11379999584621853
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,65536,256,0.011331555744012197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,64,0.002955555501911375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1024,4096,0.010839111275143094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,768,0.013788444300492605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,12288,3584,0.018397332893477548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,128,0.006198222024573221
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,8192,128,0.005063111169470681
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,6144,0.009932444327407414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,2048,0.005588444570700328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,12288,0.0199937770764033
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,512,12288,0.019280888968043856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,64,0.0059075554211934405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,2048,0.02903288933965895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,4096,65536,0.19667200247446695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,65536,0.07811555597517225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,3072,0.013816888961527081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,32,0.0026480000880029467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,2560,0.025191111697090998
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,6144,7168,0.02026844521363576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,5120,0.015666666958067153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,768,0.023810666468408372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,8192,12288,0.040018667777379356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2048,2560,0.007650666766696506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3072,2048,0.007739555504586961
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,2560,0.023532445232073467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,1536,0.008496888809733922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,7168,0.01499111122555203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,65536,0.053278221024407275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,1536,0.005515555540720622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,4096,0.008416000339719985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,1536,0.00463733325401942
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,7168,2560,0.012123555772834353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,2560,64,0.004361777669853634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,8192,0.0350195566813151
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,2048,0.02591466738118066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,3584,0.09090844127866958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,128,3584,0.01639111174477471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,128,0.028640889459186133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,65536,4096,0.14093155331081814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,3072,0.04730133215586344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,16384,1.5642364290025499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,16384,768,0.035586665074030556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,12288,0.02237866653336419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,8192,0.03178933262825012
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,2048,0.006032889087994893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,768,0.005708444449636672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,3072,0.014726221561431885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,12288,0.009204444785912832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1536,1536,0.027449778384632532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,5120,0.024085332949956257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,256,6144,0.009701333112186855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,3072,0.016910221841600206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,2048,0.00756888919406467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,256,65536,0.06844800048404269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,2048,0.008452444440788692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,512,0.005701333284378052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,5120,0.01643822259373135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,1536,0.015835555063353646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,32,0.005056888692908817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,4096,0.01719466679626041
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,1536,0.0058044443527857465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,512,0.003264888914095031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,3584,0.02204177776972453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,3584,12288,0.02070488863521152
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,128,128,0.002784888984428512
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,1536,0.005463111317820019
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,10240,0.008368888662921058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,128,0.0028408887899584244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,16384,0.023246222072177466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,4096,0.14268267154693604
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,7168,3584,0.051855110459857516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,1024,0.00851733320289188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,1536,0.009628444082207149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,5120,3072,0.20739377869500053
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,1024,0.006581333362393909
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,128,0.010245333115259806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,2560,0.008207110895050896
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,256,0.004210666649871402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3072,7168,0.014355555176734924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,768,0.004755555755562252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,12288,256,0.0035964445107513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,768,0.00553955551650789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,128,0.0027760000278552375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,1024,1536,0.01182755579551061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,1024,0.01258666647805108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,1024,0.006695111178689533
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,65536,128,0.06665866904788546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,2560,0.007606222397751278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,256,0.003867555409669876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,3584,1024,0.010760889285140567
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,1024,64,0.0052826665341854095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,12288,0.013879111243618859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,12288,0.12918933232625326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,4096,1536,0.012096888489193387
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,7168,6144,0.018505778577592637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,2048,0.06812710894478692
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,2048,0.0075662218862109715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,12288,0.027505778604083594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1024,10240,0.03539911243650649
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,7168,0.06368266873889498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,1536,16384,0.0570088889863756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,4096,0.008601778083377415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,5120,0.00870400004916721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2048,128,0.005484444399674733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,7168,0.014321777555677625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,512,0.01994311147265964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,1536,0.019732443822754752
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,5120,8192,0.02594311038653056
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,1536,16384,0.024872889121373493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,256,2048,0.012576888832781049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,3584,0.014560888210932413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,32,0.004214222232500712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,8192,0.015006222658687167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,10240,5120,0.028855111863878038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,65536,128,0.19174577130211723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,1536,0.014883556299739413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,128,0.0038248888320393036
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,2560,0.03051555487844679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,7168,0.013385777672131857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2048,3072,0.01776711146036784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,512,0.0030817778574095834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,2560,0.009504000345865885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,1024,0.00979733301533593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,1024,0.00535111129283905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,128,2560,0.006361777583758037
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,16384,0.08693244722154404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,768,0.005704000178310607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,2560,0.1346222162246704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,5120,0.019685332973798115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,128,0.007426666716734569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,10240,0.013571555415789286
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,6144,0.014104000396198697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,12288,2048,0.016473778420024447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,12288,0.01683022247420417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,2560,0.04622044497066074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,16384,1.1741608513726127
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,256,0.004914666629499859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,2560,0.014965333872371249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,2048,0.006518222391605377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,4096,0.41792088084750706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,8192,0.035016000270843506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1536,256,0.004357333398527569
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,16384,0.02699911097685496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,512,0.035848889085981585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,6144,0.011584888729784222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,768,0.005872000008821487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,3072,0.0849191082848443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,10240,128,0.0033031110134389666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3584,128,0.00627377794848548
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,10240,0.03873777720663283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,16384,2048,0.016594666573736403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,8192,0.04230133361286587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,1536,0.014257777068350049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,768,1536,0.006693333387374878
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,4096,0.025803556044896443
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,4096,0.04424533247947693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,32,3072,0.008255110846625434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,3072,0.027466666367318895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,4096,0.01812355551454756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,32,0.002809777855873108
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,5120,0.02444177700413598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,128,65536,0.011738667057620155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,512,16384,0.031122667921913996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,768,0.0038773334688610504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,12288,1024,0.014454222387737699
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,1024,0.006312888943486744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,2048,0.007615110940403408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2560,3072,0.010976889067225985
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,16384,0.0347671111424764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,3584,0.006452444526884291
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,7168,0.01872177753183577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,65536,0.24849332703484428
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,5120,0.009849777652157677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,1024,0.005150222116046482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,8192,0.02853333287768894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,12288,32,0.008463111188676622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,3072,0.008568888737095727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,256,3584,0.006694222076071634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,512,0.00869688888390859
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,12288,512,0.011136000355084738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,64,0.003040888864133093
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,5120,0.013080000049538083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,12288,0.061624891228146024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,1024,0.005355555564165115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,16384,1.6948275036282008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,1024,0.007947555846638149
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,12288,0.054393778244654335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,2560,512,0.00675733346078131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,16384,512,0.08026666773690118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,12288,8192,0.030930668115615845
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,2048,768,0.01087288889620039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,3584,0.010671111444632212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,4096,0.07087555858823988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,6144,10240,0.02348800003528595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,12288,0.007239999870459239
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,512,0.016355555918481614
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,16384,0.08921777539783055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,256,7168,0.012490666574902005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,5120,512,0.0053137776752312975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,128,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,1536,0.006882666713661618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,3584,0.007839110990365347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,256,0.01740977830357022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,1536,0.0070471109615431885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,4096,0.19604178269704184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,2048,0.04366311099794176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,2048,0.00565155554148886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,64,0.005433777968088786
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,12288,5120,0.021540444758203294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,3072,0.025133333272404138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,768,0.006052444378534953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1536,5120,0.04182222154405382
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,7168,0.01293155550956726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,8192,0.012290666500727335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2048,7168,0.07703289058473375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,2560,0.008370666868156856
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,5120,64,0.008644444247086843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,32,0.0028844444702068963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,16384,0.07172177897559272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,12288,0.026148445076412622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,16384,512,0.007443555527263218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,256,0.0041955556306574075
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,3584,0.031052443716261122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,6144,3584,0.013016000390052795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,3072,0.01648355523745219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,16384,0.03460977805985345
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,8192,0.012178666889667511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,512,0.003109333415826162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,3072,0.00868888861603207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,3072,0.013574221895800697
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,128,0.010304888917340172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,12288,12288,0.043304890394210815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,1024,0.005746666755941179
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1536,128,0.005519111123349931
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,3072,0.1112053394317627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,2560,0.012458666331238218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3072,7168,0.015104888214005364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,6144,7168,0.020602666669421725
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,2048,0.007301333049933116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,64,4096,0.009015111459626092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,12288,0.5707333352830675
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,6144,0.022805333137512207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,6144,0.059818665186564125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,10240,0.06101600329081217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,1536,0.06732088989681669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,8192,0.0203484445810318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,65536,0.08328000042173597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,128,0.0035786665976047516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,4096,7168,0.3915884494781494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,2560,0.024262222978803847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,128,6144,0.006697777658700943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2048,64,0.010124444133705562
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,64,0.005724444571468566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,6144,0.045560889773898654
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,768,0.0039813332259655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2048,10240,0.016826666063732572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,2048,0.010913777682516309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,256,0.004652444273233414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,32,5120,0.07366044653786553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,256,0.0058551111982928375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,3584,0.26068443722195095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,1024,0.00960888879166709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,32,0.0053048887186580235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,8192,10240,0.04530488782458835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,2560,0.025655110677083332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,32,0.0026168889469570587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,768,0.0047280001971456735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1536,256,0.0032382222513357797
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,4096,5120,0.01717511150572035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,10240,1.2434053421020508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,256,4096,0.011307555768224927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,5120,3584,0.024697777297761705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,12288,3072,0.017480888300471835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,12288,7168,0.03065955638885498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,5120,0.0355422231886122
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,512,0.005621333503060871
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,32,2560,0.013070222404268054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,7168,0.020853334003024634
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,1024,0.041357331805759005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,32,16384,0.008159110943476358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,5120,0.014193778236707052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,64,0.0027173333283927706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,65536,0.07708799839019775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,3072,0.021500445074505273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,7168,0.011683555940786997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,768,0.0073520003093613525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3072,16384,0.08317599693934123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,512,6144,0.008028444316652086
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,8192,0.006588444527652528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,32,0.0026648888985315957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,2048,0.03165955675972833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,4096,0.012275555895434486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,12288,0.019348444210158456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,256,0.004839999808205498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,768,5120,0.011045333411958484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,4096,0.8395413292778863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,7168,0.08707644542058308
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,3072,0.011407111254003314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,3584,0.010110222631030613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,32,0.0031191110610961914
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,8192,0.023240000009536743
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,8192,0.035395556026034884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,512,3072,0.012440000143316058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,1024,0.007640889121426477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,4096,512,0.003638222399685118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,7168,0.027504889501465693
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,512,0.04060088925891452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,512,0.008811555802822113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,512,768,0.004376888689067629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,2560,0.033415112230512835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,2560,0.00849422232972251
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,3584,0.03872444563441806
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,128,128,0.0046924443708525766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3584,768,0.006451555424266391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,6144,0.010557333628336588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,4096,0.006238222122192383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,8192,2048,0.01568888955646091
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,16384,0.01733688844574822
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,3072,0.008915555146005418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,64,12288,0.01494311127397749
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,12288,0.021202666891945735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,3584,0.07296266820695665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,4096,0.026335999369621277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,12288,10240,0.34537243843078613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,32,0.007857778006129796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,32,64,0.008396444221337637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,7168,0.01238666640387641
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,3584,0.08122489187452528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,64,0.0028764444092909494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,1536,0.00833511104186376
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,12288,0.018812444474962022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2560,5120,0.01107377807299296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,8192,5120,0.02296177711751726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,4096,0.2734266652001275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,10240,512,0.01683200067943997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,768,4096,0.008242666721343994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,256,0.006990222467316522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,1536,8192,0.01407199932469262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,2560,0.05056711037953695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,7168,0.08915288580788507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,512,3072,0.006945778098371293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,3584,8192,0.015448888142903646
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,512,0.004898666507667965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,3584,0.011183111204041375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3584,3584,0.030877331892649334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,3584,0.01038933296998342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,16384,8192,0.05761688947677612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,5120,768,0.00977244476477305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,2048,0.018365333477656048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,128,0.004193777839342753
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,768,0.0038213332494099936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,3584,0.12323110633426243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,10240,2560,0.017097777790493436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,1024,0.005663110978073544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,7168,0.01791200041770935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2560,4096,0.0124177775449223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,64,768,0.00369422220521503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,4096,256,0.018425777554512024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,128,768,0.003552888830502828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,7168,0.035975999302334256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,6144,128,0.0053084443012873335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,5120,768,0.004834666848182678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,1024,7168,0.01332355539004008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2048,8192,0.014304889572991265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,128,0.004287111262480418
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,6144,0.14860799577501085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,16384,0.026688888669013977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,1536,0.006712889091836081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,5120,0.02257866660753886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,6144,1024,0.028636445601781208
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,4096,0.1579315529929267
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,32,12288,0.05310577816433377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,3072,64,0.005294222384691238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,1536,2.556004418267144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,128,0.004404444247484207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2048,2560,0.00925333301226298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,64,0.003646222253640493
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3072,4096,0.012496000362767113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,768,0.006320888797442119
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,128,0.0046160001721647055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,12288,0.18485600418514678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,2560,0.019536889261669584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,1024,0.007714666426181793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1024,3584,0.007927999728255803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,512,3584,0.008188444707128737
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,3584,128,0.004559111264016894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,12288,0.010414222048388587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,2048,0.008421333299742805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,4096,0.06212888823615181
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,32,256,0.0030782222747802734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,512,0.004577777865860197
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,3072,2048,0.00777422222826216
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3584,2560,0.008083555433485243
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,512,8192,0.031685332457224526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,10240,0.043024000194337636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,32,0.005470222069157495
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,8192,65536,0.2210986614227295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,2048,0.008191111187140146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,2560,0.006678222368160884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,3584,0.013600889179441663
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,2560,0.007072889142566257
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,6144,0.5827813148498535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,1536,0.01073066641887029
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,12288,3072,0.026458667384253606
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,768,2048,0.006846222198671765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,256,0.003310222178697586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,3072,3072,0.010213333699438307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,2560,10240,0.3913066652086046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,12288,0.05311200022697449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,512,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,16384,0.02493511140346527
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,4096,3584,0.06386133035024007
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,64,0.012440000143316058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2048,2560,0.017072889539930556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,64,768,0.006177777631415262
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,7168,2048,0.01127022256453832
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,4096,0.01309333327743742
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,3072,0.012208888928095499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,4096,12288,0.06494488981035021
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,5120,0.009468444519572789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,2048,0.08283555507659912
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2560,2048,0.00703288863102595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,12288,1.3300488789876301
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,7168,3072,0.015591111448076038
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,8192,2048,0.015160888433456421
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,32,0.011208000282446543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,32,0.0063697778516345555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,65536,0.10556355449888442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2048,4096,0.021907554732428655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,65536,768,0.041999111572901406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,128,0.004876444323195351
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,7168,256,0.0120328888297081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,16384,0.05267555514971415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,2048,5120,0.012693333129088083
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,16384,65536,0.3257591194576687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,12288,0.016540444559521146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,1536,0.014496000276671516
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,256,0.0070017774899800616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,768,0.026512889398468867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,128,256,0.002874666617976295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,3584,0.01238933371173011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,3584,0.10256266593933105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,10240,0.012044444680213928
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,5120,0.032320890161726214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,7168,0.0081386665503184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,5120,32,0.005323555734422471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,1024,0.20846311251322427
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,8192,512,0.0058088890380329555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,8192,1536,0.016564443707466125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,7168,0.009465778039561378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,64,0.00518133325709237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,2560,0.007640000018808577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,768,0.003605333467324575
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,10240,3584,0.04169955518510607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,4096,0.016757332616382174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,64,0.007128888534175024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,64,0.0052471111218134565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2560,768,0.005702222386995952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,2048,0.011320888996124268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,128,0.0029004443850782183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,2048,3072,0.008612444831265343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,12288,4096,0.039919111463758684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,2560,0.037273777855767146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,1536,0.006325333482689328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,4096,2560,0.010064000056849586
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,1024,12288,0.012306667036480375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,8192,3584,0.016819554898473952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,10240,0.03658933440844218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,512,8192,0.009747555686367882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,3584,0.7487849129570855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,512,0.02462844385041131
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,3072,0.031002667215135362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,32,0.003343111111058129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,256,12288,0.015442666080262927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,6144,128,0.007699555820888943
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,10240,3072,0.01389955480893453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,10240,65536,2.154209772745768
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,8192,0.018762666318151686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,512,0.009081777599122789
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,768,0.006511111226346757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3072,512,0.003779555360476176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3072,65536,0.2889964315626356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,1024,0.01832266648610433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,5120,0.013655111193656921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,12288,7168,0.0314702226055993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,16384,0.3479964468214247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,4096,0.012095111111799875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,65536,2560,0.05282844437493218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,3584,0.03771377934349908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,6144,512,0.005613333235184352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,2560,7168,0.014865777558750577
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,3584,0.009297777381208207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1024,7168,0.011725333001878528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,3584,0.0075804442167282104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,4096,12288,0.025983999172846477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,3072,0.006907555378145642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,64,0.006430222342411677
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,65536,0.10040711032019721
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,2048,512,0.005049777941571342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,256,0.006151999864313338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,32,64,0.005229333208666907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,768,8192,0.013435555828942193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,3072,0.05387466814782885
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,12288,1024,0.027373333772023518
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,10240,0.2527528868781196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,65536,1536,0.04904177784919739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,256,512,0.0032675556010670135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,6144,0.009132444030708736
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,128,0.009604444106419882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,8192,0.5622853173149956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1024,8192,0.015478221906556023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,12288,128,0.004229333251714706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,4096,3072,0.13971643977695042
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2048,128,0.005075555708673266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,64,768,0.012896888785892062
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,256,0.005192888693677055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,16384,0.0391937792301178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,65536,0.07736800114313762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,5120,0.4810933536953396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2048,768,0.005569777968857024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,1024,0.00545155546731419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,1024,512,0.01240711079703437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,256,768,0.008202667037645975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,5120,6144,0.3359751171535916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,16384,0.02546577817863888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,1536,7168,0.011843555503421359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,7168,0.02237066626548767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,65536,12288,0.40235021379258895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,6144,2048,0.16546221574147543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,65536,0.2028728855980767
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,10240,0.013869333598348828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,128,8192,0.01129333343770769
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,256,3584,0.007353777686754863
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,128,0.00996888925631841
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,768,0.005769777629110549
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,1024,1536,0.006872889068391588
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,1024,0.00425155543618732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3584,65536,1.8842088911268446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,8192,10240,0.9297617806328667
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,12288,0.055941330062018506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,5120,16384,0.1441991064283583
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,5120,2048,0.009406222237481011
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,3584,0.008737777670224508
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,5120,1536,0.11796800295511882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,1536,0.00489688871635331
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,768,12288,0.02750844425625271
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,3072,64,0.006615110983451207
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,12288,0.14842844009399414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,7168,0.2543768882751465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,12288,1536,0.02307022280163235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,4096,0.03529777791765001
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3584,768,0.008436444732877944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,1024,0.004547555413511065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,256,2048,0.005702222386995952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,16384,256,0.004968889057636261
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,5120,0.008961777720186446
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,10240,16384,0.04878044459554884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,7168,0.029007110330793593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,512,8192,0.00885333369175593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3072,2560,0.008803555534945594
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,4096,0.014635556273990206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,6144,0.0798542234632704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,128,10240,0.03731822305255466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,8192,128,0.004187555362780889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,12288,128,0.006690666493442323
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,6144,0.03304444419013129
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,256,0.00518488883972168
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,7168,0.10907999674479167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,128,0.007059555914666917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,256,0.011598221957683563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,16384,7168,0.04092622134420607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,10240,256,0.015110222829712762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,16384,128,0.009525333013799455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1536,12288,0.02790044413672553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,12288,12288,0.07528711027569242
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,10240,7168,0.04393066631423103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,32,768,0.0035911111368073356
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2560,1024,0.004010666575696734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,6144,16384,0.08439289198981391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,12288,0.014013333453072442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,12288,0.02190133266978794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,1024,3072,0.006612444503439798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,2048,12288,0.04860622353023953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,10240,3072,0.014176888598336114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,1536,0.05394311083687676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2048,768,0.005786666853560342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,6144,0.01640977793269687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,6144,0.013001778059535556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,5120,3072,0.010974222587214576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,12288,0.018593778212865193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,128,0.004974222017659081
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,768,0.008221333225568136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,768,0.005764444255166584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1024,65536,0.07347022162543403
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,768,0.010392000277837118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,512,0.003519111209445529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,5120,0.07011644707785712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,7168,65536,0.15812710920969644
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,10240,0.015078221758206686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,32,12288,0.015554666519165039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,512,3072,0.0073928890956772705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,16384,4096,0.026100445124838088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,2560,256,0.0032799999333090256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,7168,12288,0.04522488845719231
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,10240,0.05925510989295112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,65536,768,0.026498667067951624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,1024,0.00573511090543535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,2560,0.006001777946949005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,3072,0.021183111601405676
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,768,65536,0.0658586687511868
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,2048,0.04253600041071574
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,5120,1024,0.009806222385830348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,7168,5120,0.11232088671790229
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3584,6144,0.48267290327284074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,1536,6144,0.011151110960377587
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,3072,0.2875644365946452
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1024,65536,0.04155288802252875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,65536,0.7775875727335612
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,512,0.00434933313065105
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,5120,0.01036266651418474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,32,0.004574222283230888
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,256,0.0029404444826973807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,5120,1024,0.008242666721343994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,512,256,0.0034968890249729156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,128,65536,0.014700444208251106
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,512,4096,0.010292444792058732
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,2048,65536,0.08829422129525079
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,16384,6144,0.0517155561182234
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,768,3584,0.06318489048216078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,12288,0.01906844476858775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,128,0.0038906666967603895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,1536,4096,0.008458666503429413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,1536,0.012506666282812754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,8192,0.04581244455443489
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,768,0.0070462218589252895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,64,0.004969777746333016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,16384,0.0706497761938307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,2048,0.005893333504597346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,2560,0.01479822231663598
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,7168,64,0.012763555679056378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,64,0.006700444552633498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3072,2560,0.010315555665228102
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,128,0.010584889186753167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,3072,0.012472888661755456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,6144,3072,0.01316088851955202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,4096,0.01331911070479287
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,512,0.007259555160999298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,768,256,0.004381333374314838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,2560,0.01090222183201048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,7168,1536,0.008552889029184977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,12288,256,0.009205333060688442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,512,128,0.004309333446953031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,512,16384,0.020418667131000094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,65536,0.10274222162034775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,256,0.00434400017062823
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,5120,0.01405066748460134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,128,0.05351644423272875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,12288,0.016763556334707472
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,10240,16384,0.0948080023129781
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,512,8192,0.008257778154479133
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,1536,0.00703111125363244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,10240,128,0.004386666748258802
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3584,6144,0.018560000591807894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,64,1536,0.0046471113132105935
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,512,16384,0.07526044713126288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,6144,256,0.04136088821623061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,16384,0.09624355369144016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,12288,6144,0.04053599966896905
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,2560,0.010616000327799054
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,768,0.007909333540333642
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,512,10240,0.015855999456511605
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,65536,4096,0.14352799786461726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,6144,2560,0.012152888708644442
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,16384,2048,0.01573511130279965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,10240,2048,0.015456888410780164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,16384,768,0.09627999862035115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,64,1024,0.0038480001191298165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,8192,16384,0.19423466258578828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,256,1536,0.004663111021121343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,1024,0.010562666588359408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,12288,0.007506666911972894
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,1536,0.0063484443558586975
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,1024,12288,0.061315556367238365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,4096,3072,0.01609777741962009
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2560,768,0.004920000003443824
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3584,512,0.005562666803598404
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,10240,0.016582222448454965
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,7168,12288,0.031331555710898504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,8192,64,0.005044444567627377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,8192,0.011574221981896294
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,256,0.004557333472702238
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,128,1024,0.004017777740955353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,128,6144,0.010075555907355415
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,4096,0.017254221770498488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,65536,0.25685156716240776
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1024,1536,0.007113777928882175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,7168,32,0.005555555638339784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,2560,256,0.004120888809363048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,3072,0.01906399925549825
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,1024,0.016501333978441026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,256,2048,0.006296000132958095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,32,3072,0.007366221812036302
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,2048,0.0058604445722368025
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,65536,8192,3.281346638997396
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,65536,7168,0.2211973269780477
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,4096,256,0.009975999593734741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,16384,65536,0.2805573410458035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,7168,0.023267555567953322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,7168,3584,0.014762666490342883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,768,1024,0.01000533335738712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3584,128,0.004044444610675176
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,7168,128,0.006320000108745363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,128,4096,0.0063457778758472866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,65536,3072,0.5797911220126682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,16384,0.027054222093688116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,3072,0.011515555282433828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,6144,256,0.011770666473441653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,8192,0.05576088693406847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,2048,0.16094222333696154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,4096,2560,0.008638222184446123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,3072,32,0.02966488732231988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,3072,0.026944888962639704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,128,0.0038240001433425476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,512,1024,0.01538222200340695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,32,0.0025742221623659134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,5120,0.49963998794555664
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,8192,2560,0.014256889621416727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,2560,0.024314666787783306
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,10240,1024,0.008560889297061497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,256,0.0037377778854635027
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,16384,0.08994311094284058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,1024,0.04678844412167867
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,8192,0.03165866600142585
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,512,10240,0.014474666780895658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,3584,0.24219822883605957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,4096,768,0.004888888862397936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,5120,0.16002578205532497
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,3584,0.010051555931568146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,128,0.005106666849719154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,256,768,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,8192,0.04010933306482103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,3072,3584,0.010888888604111142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,4096,0.017417778571446735
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,1024,0.006178666734033161
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1536,65536,0.08922488821877374
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,5120,0.17669333351982963
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1536,4096,0.008479110896587372
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,7168,0.01498311095767551
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,4096,0.13887733883327907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,256,512,0.007157333195209503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,6144,1024,0.048819555176628955
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,5120,0.018008000320858426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,8192,12288,0.05658311314053006
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,2560,64,0.008439111212889353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,1536,0.010676444404655032
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,32,2560,0.007101332975758447
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,8192,3584,0.08955378002590603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1024,2048,0.04788444439570109
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,2048,0.006960888703664144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,128,2048,0.006633777585294511
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1536,16384,0.022057778305477563
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,4096,4096,0.011780444118711682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,3072,0.07097422414355807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,128,512,0.004098666624890434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,3584,16384,0.03355733222431607
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,2048,0.024139554964171514
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,5120,10240,0.04068355427847968
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,5120,0.023215999205907185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,512,0.007300444775157505
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,1536,64,0.005405333307054307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,3584,0.006264888991912206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,768,0.005701333284378052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,32,0.005383111122581694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,7168,1536,0.008715555899673039
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,768,1536,0.033626665671666466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,32,0.0046364445653226645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,2560,0.00795911086930169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,8192,5120,0.032623999648623996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,7168,0.02660977840423584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,5120,0.009829333258999718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1536,256,0.01682933337158627
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,256,0.006701333241330252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,512,0.005620444400442972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,4096,3584,0.013329777452680798
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,5120,0.0075706665714581805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,2560,0.013437333206335703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,16384,3072,0.02812622321976556
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,6144,7168,0.02527911133236355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,2560,0.005605333381228977
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,64,7168,0.010959111154079437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,3072,0.010261333651012844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,8192,2.0134080251057944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,6144,0.01633777717749278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,16384,12288,0.0579706695344713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,8192,5120,0.021703999903466966
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,8192,768,0.006222222414281633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,3072,2560,0.12071288956536187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,768,0.004979555391603046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,1536,2048,0.008418666819731394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,5120,65536,0.11719732814364964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,7168,32,0.0069715554515520734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,12288,12288,0.06863288746939765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,3584,3072,0.0177528891298506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1536,10240,0.17199022240108916
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,32,1536,0.0046942221621672315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,6144,4096,0.015509333875444202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,64,0.0029315555261241067
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,10240,0.01347644461525811
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,6144,10240,0.036920001109441124
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,64,0.002586666701568498
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,8192,0.09137599998050266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,6144,0.024157333705160353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2048,12288,0.03472088774045309
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,65536,512,0.06878133614857991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,6144,128,0.005323555734422471
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,12288,3072,0.017502221796247695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,32,2048,0.005416888743638992
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,2048,0.006713777780532837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,1536,0.06386311186684503
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,2048,0.010467555787828235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,3584,10240,0.02903555499182807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,256,256,0.004229333251714706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,10240,1024,0.011815999945004782
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,65536,0.37733154826694065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,8192,0.024735111329290602
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,256,3584,0.00777333312564426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,512,5120,0.01689511040846507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2048,1024,0.006484444356626934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,7168,7168,0.04006933172543844
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,65536,3.512153837415907
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,2560,3584,0.012520888613329994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,4096,0.0114702218108707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,512,0.004278222305907143
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,512,768,0.005780444376998478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,6144,1024,0.010125333236323463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,6144,0.013316444224781461
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,128,10240,0.016140444411171805
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,1536,0.0046311111913787
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,256,6144,0.008432000047630733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,128,8192,0.011764444410800934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,12288,0.026868444350030687
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1536,6144,0.013589333328935834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,8192,0.07713155614005195
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,768,32,0.002675555646419525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,3584,768,0.003972444269392225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,7168,1536,0.010947555303573608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,6144,64,0.0029395555870400537
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,8192,0.14088533984290227
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,128,0.003177777760558658
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,10240,0.016765332884258695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,512,3072,0.007108444141017065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,5120,256,0.006539555473460092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,32,0.005151111218664381
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,768,1024,0.004960000101062986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,65536,768,0.02164177762137519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,16384,0.021576000584496394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,1024,0.006911110960774952
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,3072,1024,0.012445333103338877
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,16384,6144,0.05020977722273933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,8192,7168,0.09208533498975967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2560,64,0.003220444545149803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,128,0.004030222280157937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,1024,0.049606220589743726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,7168,1536,0.008638222184446123
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1024,65536,0.0418817765182919
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,5120,32,0.002733333243264092
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1536,8192,0.030014223522610132
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,768,64,0.0064560001095136
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,8192,512,0.007258666886223688
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,5120,4096,0.03174755639500088
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,12288,0.0236826671494378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,8192,0.006254222244024277
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,128,0.004252444538805219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,6144,16384,0.04818755719396803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,10240,3072,0.01566844516330295
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,6144,128,0.004300444490379757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,128,6144,0.026533333791626826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,12288,0.04055822226736281
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,3584,0.010013333625263637
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,16384,5120,0.026555554734336004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,4096,8192,0.021672000487645466
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,3072,0.005872000008821487
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,12288,256,0.009400000174840292
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,5120,0.022996443841192458
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,1536,12288,0.01662488944000668
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,64,2048,0.0053395554423332214
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,65536,768,0.03273955649799771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,10240,65536,0.2706693278418647
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,6144,6144,0.018586667047606576
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,10240,10240,0.05968622366587321
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,2560,5120,0.014330666926172046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3584,2048,0.012347555822796293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,8192,2560,0.13432266977098253
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,10240,0.3241688940260145
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,8192,0.01109866632355584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,7168,5120,0.0354746679464976
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,64,12288,0.041307555304633245
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,8192,10240,0.053767111566331655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,768,128,0.003863999827040566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,12288,6144,0.055777774916754834
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,6144,0.006193777753247156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,2560,0.05607733461591932
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,256,3584,0.008728889127572378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,3584,0.011206222077210745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1024,512,0.00332622230052948
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,16384,0.022642667094866436
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,2048,1536,0.006320000108745363
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,65536,17.03729078504774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3072,128,0.00518844442235099
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,8192,64,0.01370044466522005
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,4096,4096,0.010405333505736457
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,10240,0.021427555216683283
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,3072,6144,0.0163102216190762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3072,5120,0.012568888564904531
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,2048,0.01532444523440467
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,256,4096,0.007671111159854465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1536,64,0.005443555613358815
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,1536,768,0.0038444445365005066
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,7168,128,0.004640000147951974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,3072,3072,0.012930666406949362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,10240,0.033773332834243774
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,1024,0.0226239992512597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,6144,1536,0.009169778062237633
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,5120,0.01737777723206414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2560,6144,0.02443644404411316
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,2560,256,0.006457777900828256
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,65536,1536,0.06987733311123319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,1536,0.012254222399658628
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,256,0.00573244442542394
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1024,128,0.0028444443725877335
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3584,512,0.004909333255555895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,64,2048,0.0071928890215026
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,7168,3584,0.012631110846996307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,2048,8192,0.03997333182228936
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,1024,0.005124444348944558
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,128,6144,0.005868444426192178
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,65536,7168,11.39640892876519
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2560,1536,0.00812888890504837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,4096,0.08264266782336764
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,16384,3584,0.08205066786872016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,12288,6144,0.03560533457332187
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,1536,7168,0.014567111929257711
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,4096,0.040960888067881264
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,512,2560,0.011659555964999728
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,1536,0.006680888848172293
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,2048,128,0.007857778006129796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2048,1536,0.006303999986913469
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,2048,1024,0.003996444245179494
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,512,10240,0.023016000787417095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,2560,0.02232088810867733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,64,12288,0.007630222373538547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,8192,0.016182222300105624
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,65536,3584,0.061976889769236244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1536,8192,0.013370667066839008
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,16384,0.04871822065777249
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3584,2560,0.008681777450773451
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,1024,512,0.005139555368158552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,2560,7168,0.013510222236315409
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,6144,65536,0.6513991355895996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,4096,12288,0.03958755400445726
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2048,768,0.006786666810512543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,8192,7168,0.054423110352622144
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,1024,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,3584,768,0.0058311112225055695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,5120,0.012647999657524956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,65536,7168,0.14183199405670166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,12288,0.04051466782887777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,1536,4096,0.021301334102948506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,768,16384,0.012467555701732635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,65536,64,0.4247715473175049
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3072,1536,0.007311999797821045
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,1024,128,0.0038968887594011095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,10240,2560,0.2946231100294325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,1024,128,0.003346666693687439
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,3072,0.28675733672247994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,256,0.003601777884695265
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,10240,1024,0.00979288915793101
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,16384,10240,0.23997333314683703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,3072,3072,0.009389333426952362
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,3584,0.06422311067581177
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,3072,256,0.0034888887570963967
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,7168,128,0.007328888608349695
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,10240,0.04201599955558777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,12288,512,0.010285333626800114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,256,0.005958222266700532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,3584,128,0.007239110767841339
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,256,2048,0.015674667225943673
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,7168,5.780866834852431
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,65536,5120,0.14504888322618273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,8192,0.023955555425749883
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,1536,0.0068897778789202375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,8192,0.026171555121739704
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,8192,7168,0.037462221251593694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1024,256,0.01548266741964552
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,768,6144,0.010199999643696679
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,3072,0.013832000394662222
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,64,768,0.003383999897374047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,3584,0.009573332965373993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,3584,4096,0.011208888557222156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,64,2048,0.004976888911591636
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,7168,1024,0.01090755561987559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,8192,1024,0.007738666401969061
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,128,0.0028311111446883944
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,65536,256,0.012233778006500669
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,4096,0.02102577851878272
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,1536,768,0.0041582224269707995
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,3072,512,0.0038124442928367188
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,12288,10240,0.09691111246744792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,6144,0.04397955536842346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3584,65536,0.12201155556572808
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,2048,768,0.041894220643573336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,65536,3.2705927954779734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,65536,6144,0.2262942261166043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,16384,5120,0.055712891949547656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,128,0.007342221836249034
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,10240,5120,0.036898665957980685
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,2048,0.025106666816605464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,4096,64,0.005743999862008625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,768,2048,0.006126222097211414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,12288,7168,0.0462355547481113
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,256,3584,0.010112889111042023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3072,256,0.003087999920050303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,65536,3072,2.516140407986111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,16384,0.1987253295050727
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,2560,5120,0.026541334059503343
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,4096,1536,0.007890666524569193
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,6144,0.013481777575280933
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,256,64,0.0031031111462248694
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,768,256,0.006028444402747684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,3584,0.010771555205186209
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,768,16384,0.06198222107357449
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,1536,5120,0.16797510782877603
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,128,7168,0.011663999822404651
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,5120,65536,0.09806133641137017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,65536,0.8578622076246473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,4096,0.00905511114332411
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,16384,768,0.03310666812790765
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,512,4096,0.006156444549560547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,4096,8192,0.01573244399494595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,768,16384,0.013527111046844058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,8192,0.008651555412345463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,5120,16384,0.062126219272613525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,1024,2048,0.01419911119672987
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,2048,768,0.004959110998445087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,12288,0.03375644485155741
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,65536,3072,0.06279555294248793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,128,10240,0.020709334148301017
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,3584,6144,0.01870577699608273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,256,128,0.004537777768241035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2048,32,0.0031146665828095544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,16384,10240,1.0428187052408855
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,10240,65536,0.410720878177219
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,5120,3072,0.03276266654332479
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,256,65536,0.41841332117716473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,5120,6144,0.02656800051530202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,128,2048,0.005786666853560342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2048,7168,0.013562666873137156
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,4096,0.009504888620641496
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,12288,10240,0.8316782315572103
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,3072,0.012599999705950419
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,64,8192,0.01277955538696713
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,4096,65536,0.08797333637873332
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,7168,5120,0.018290667070282828
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,32,0.0032888888898822996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,7168,8192,0.806853347354465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,768,0.005092444519201915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,64,2048,0.012146666646003723
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,1536,8192,0.011615110768212212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,10240,0.04203822215398153
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,1536,512,0.005575110928879843
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2560,32,0.005932444499598608
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,32,1536,0.016178665889634024
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,2560,512,0.0035582222044467926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,256,65536,0.06639466683069865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3584,6144,0.04247200157907274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,5120,1024,0.006030222194062338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,512,2048,0.006243555496136348
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,16384,64,0.1142888863881429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,1536,256,0.00601244428091579
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,6144,5120,0.023048889305856492
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,768,2048,0.007647111184067196
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,4096,0.25619999567667645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,768,3072,0.007652444144090016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,32,0.007605333295133378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,32,0.0031999999450312722
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,128,2560,0.006319111006127463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,7168,0.017078222499953378
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,6144,0.006461333483457565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,3072,0.0069146665434042616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,10240,3584,0.0162444445821974
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,5120,768,0.010649777948856354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,7168,3072,0.03878666626082526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,64,64,0.002518222149875429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,16384,7168,0.0806444419754876
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,32,3072,0.0075626663035816616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,5120,65536,2.3256568908691406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,7168,0.050473776128556996
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,12288,8192,0.12265777587890625
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,1024,0.0047271110945277745
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,2048,2048,0.006512889017661412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,512,0.004678222040335338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,16384,16384,0.7149706416659884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,32,1024,0.004101333518822988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,2560,8192,0.052256888813442655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,768,65536,0.329365332921346
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,3072,512,0.005941333456171884
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,128,2560,0.008119111259778341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,1536,5120,0.012196444802814059
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,7168,0.07179555628034803
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1536,8192,0.01385600037044949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,8192,7168,0.7314675649007162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,5120,0.0075795559419526
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,7168,768,0.007625777688291337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,10240,5120,0.024922667278183833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,32,256,0.0031946665710873077
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,2048,16384,0.02311466634273529
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,5120,7168,0.025711999999152288
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,512,0.006583999842405319
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,10240,3584,0.04865333437919617
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,768,2048,0.008267555799749162
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,1024,0.00906755526860555
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,1536,0.016805332567956712
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,8192,2560,0.018764444523387484
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3072,8192,0.06872177786297269
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,256,0.0041031113101376426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,5120,3072,0.010900444454616971
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,2048,128,0.003489777859714296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,8192,7168,0.03069066670205858
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,64,12288,0.032123555739720665
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,12288,0.08051111300786336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,128,0.003917333152559069
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,4096,3584,0.01074577785200543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,6144,0.027476444840431213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3584,12288,0.03233155608177185
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,7168,16384,1.5595128801133897
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,1024,0.0047084444926844705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,7168,0.02865333358446757
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,8192,16384,0.06080177757475111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,128,0.002785777880085839
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,16384,8192,0.06250044372346666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,512,4096,0.006150222072998683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,3584,5120,0.0536426670021481
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,3584,5120,0.013188444077968597
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,4096,0.15178844663831922
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,16384,1024,0.015604444675975375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,6144,8192,0.02764355474048191
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,256,0.005096000101831225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,8192,0.01367644468943278
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,4096,12288,0.043699555926852755
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,128,0.005320888840489917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,5120,5120,0.023823999696307715
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,512,2048,0.005304000029961268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,64,12288,0.015213333898120456
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,1536,0.006447111152940326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,5120,2048,0.0087333329849773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,1536,0.013411555853154926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,7168,0.026936888694763184
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,10240,4096,0.1272719966040717
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,5120,3072,0.010447111394670276
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,16384,16384,3.371895896063911
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,6144,12288,0.04576622115241157
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3072,6144,0.03238488899336921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,3072,256,0.004889777965015835
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,10240,3584,0.10343555609385173
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,1536,1536,0.005031110925806893
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,8192,4096,0.02881066666709052
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,10240,0.013419555293189155
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,512,8192,0.04933422141604953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,1024,768,0.005786666853560342
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,512,3072,0.026678222748968337
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,128,2560,0.0064239998658498125
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,12288,4096,0.1479297743903266
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,4096,2560,0.009665778113736046
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3072,7168,0.23773601320054796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,128,1024,0.011277333729796939
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,65536,768,0.02935111191537645
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,7168,128,0.003379555626047982
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2048,8192,0.015858666764365304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,4096,64,0.0027315556589100095
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,6144,1536,0.1079288919766744
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,3072,4096,0.08787111441294353
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,7168,128,0.005669333454635408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,768,2048,0.007043555378913879
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2048,3584,0.0796639985508389
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,6144,1024,0.006538666784763336
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,65536,0.07972800069385104
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,128,10240,0.018043556147151522
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2560,12288,0.2262942261166043
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,2560,0.013633777697881063
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,256,12288,0.01918044520748986
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,65536,4096,0.3684711191389296
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,16384,128,0.01351999988158544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,256,2560,0.0064204442832205035
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,2560,0.005533333453867171
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,3072,65536,0.10467733277214898
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,65536,0.0677315526538425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,1024,6144,0.00960977789428499
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,8192,0.021789333886570398
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,4096,6144,0.013849777479966482
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,1536,0.013111111190583972
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,3072,32,0.006660444455014334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,65536,3072,0.11986755000220405
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,1024,1536,0.0058373332851462895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,65536,8192,0.3829190995958116
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,6144,3072,0.013721778160995908
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,6144,6144,0.031006223625606958
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,5120,0.09806399875217015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,5120,10240,0.10822044478522407
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,8192,1.7155413097805445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,2560,8192,0.257422235276964
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2048,8192,0.016890666551060148
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,1536,8192,0.032456000645955406
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,1024,1024,0.005640000104904175
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,32,0.005584888988071018
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,16384,0.0244195560614268
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,4096,16384,0.1347173319922553
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,128,2048,0.005096000101831225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3584,3072,0.047727111313078135
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,5120,32,0.030840890275107488
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,256,0.0046924443708525766
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,8192,512,0.005161777966552311
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,6144,128,0.020960888928837244
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,2560,32,0.01148444414138794
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,2560,16384,0.023789332972632513
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,1024,3584,0.00868000007337994
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,32,512,0.0034808889031410217
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,2048,7168,0.024430221981472437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,2048,0.01458666721979777
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,5120,0.022616000639067754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,1024,3072,0.0069466663731469055
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,16384,0.09616266356574164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,5120,0.018057778477668762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,256,4096,0.008714666797055138
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,10240,2560,0.012416000167528788
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,65536,2560,0.12516889307234022
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,16384,6144,0.07645155323876275
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,3584,6144,0.02453511125511593
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,16384,0.07517511314815946
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,65536,0.015056888262430826
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,65536,0.134479112095303
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,1536,16384,0.023807111713621352
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,2048,7168,0.1146586603588528
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,16384,10240,0.04651288853751289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2048,7168,0.060848891735076904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,16384,2048,0.013696889082590738
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,512,10240,0.01462044484085507
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,4096,512,0.025568000144428674
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,3584,0.015273777974976433
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,6144,32,0.006701333241330252
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,256,2048,0.006002666635645761
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,32,0.0034417777011791864
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,4096,0.018114666144053142
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,5120,32,0.006653333289755716
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,4096,3584,0.012125333150227865
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,2560,3584,0.052192889981799655
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,3072,16384,0.17766577667660186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,65536,1024,0.0640231106016371
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,768,4096,0.008917333351241218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,3072,10240,0.023592000206311543
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,1024,32,0.0029102222373088202
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,7168,3072,0.011372444530328115
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,3072,65536,0.11491200659010146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,8192,1536,0.009933333430025313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,16384,256,0.007057777709431118
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,12288,5120,0.023969777756267126
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,3072,12288,0.07858310805426703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,7168,12288,0.045032888650894165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,4096,2048,0.011759111450778114
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,512,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,1536,256,0.004273777620659934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,6144,7168,0.018047110901938546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,7168,128,0.007641777396202087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,2560,32,0.0029306666304667792
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,768,0.013190222283204397
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,1024,128,0.005405333307054307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,512,128,0.010135999984211391
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,768,128,0.008557333714432186
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,768,5120,0.014030221435758801
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,256,6144,0.016201777590645682
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,1536,512,0.0033555556502607134
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,1024,5120,0.009509333305888707
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,256,256,0.004033777862787247
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,10240,128,0.06175021992789375
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,1024,0.004592000196377437
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,8192,0.014937778313954672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,6144,0.07860266500049166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1536,65536,0.06812089019351535
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,7168,16384,0.1125368939505683
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,1536,0.009005332986513773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,64,16384,0.008195555872387355
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,1024,12288,0.132313781314426
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,64,10240,0.008977777428097194
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,2560,10240,0.016543999314308167
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,8192,16384,0.3580542140536838
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,128,32,0.0026142222599850762
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,1536,512,0.00351555562681622
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,2560,768,0.006383111079533895
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,256,1024,0.005143110950787862
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,32,5120,0.0063075555695427795
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,768,7168,0.011688888900809817
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,2560,12288,0.020807110601001315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,8192,256,0.0075493330756823225
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,256,12288,0.0068915556702348925
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,6144,3584,0.03331733412212796
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,8192,1024,0.009322666459613377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,7168,128,0.005496888938877318
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,64,512,0.00313688897424274
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,10240,0.02509511179394192
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,2560,5120,0.11303555965423584
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,1536,3072,0.013157332936922709
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,10240,0.03554577628771464
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,10240,768,0.008729777402347988
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,64,5120,0.006558222075303395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,65536,1536,0.030551112360424463
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,5120,4096,0.03031555480427212
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,3072,512,0.004822222308980094
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,2560,4096,0.012085333466529846
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,4096,256,0.005331555588377847
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1024,4096,65536,0.42598311106363934
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,2048,0.04275822308328417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,8192,256,0.007963555554548899
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,16384,2048,0.019267555740144517
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,12288,16384,0.10339288579093085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,256,10240,0.009901333186361525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,6144,32,0.006024000131421619
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,7168,4096,0.10795555512110393
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,16384,1024,0.06244888570573595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,512,0.005527110977305307
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,5120,3584,0.022428444690174524
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,3072,256,0.015971556305885315
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,6144,1024,0.010209778116808997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,12288,2048,0.018224888377719455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,2048,16384,0.02920000089539422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,1024,0.011340444286664328
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,65536,12288,0.5409484439425999
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,128,2560,0.0074728892909155945
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,3072,3072,0.20351289378272164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,512,256,0.005389333185222413
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,65536,5120,4.335391998291016
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,65536,5120,0.09563733471764459
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,2560,768,0.0041848888827694785
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,7168,8192,0.022358222140206233
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,7168,3072,0.049591110812293164
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,16384,0.053901334603627525
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,256,64,0.005469333380460739
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,256,128,0.00267999991774559
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,7168,7168,0.049526221222347684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,16384,128,0.006888888776302338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,256,6144,0.009654222263230218
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,65536,0.0668213301234775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,128,16384,0.01034311122364468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,3072,1024,0.005390222287840313
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,2560,7168,0.021337777376174927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,4096,3072,0.009376000199053023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,1024,1024,0.005254222287072076
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,12288,512,0.00517511119445165
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,8192,1024,0.007526222202512953
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,64,256,0.0029324444217814338
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,4096,0.1284488836924235
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,16384,1024,0.011857777833938599
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,16384,512,0.15146666102939182
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,16384,0.15405600600772434
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,3584,32,0.005883555445406172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,768,4096,0.009493333597977957
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,64,3584,0.021216000119845074
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,32,1024,0.003959111041492886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,64,10240,0.013535111314720578
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,5120,0.01919822229279412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,16384,0.1927635537253486
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,256,10240,0.006540444576077991
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,128,0.0034373334298531213
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,65536,5120,0.08937955564922756
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,2560,5120,0.19184711244371203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,65536,32,0.0041964443193541635
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,32,6144,0.009992889232105678
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1024,65536,2048,0.20921688609653047
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,1536,0.005950222412745158
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,1536,12288,0.03964089022742377
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,64,32,3072,0.006149333384301927
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,5120,2560,0.010574222438865237
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,768,8192,0.013109332985348172
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,16384,768,0.00998577806684706
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,10240,3584,0.02278933260175917
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,1536,256,0.011843555503421359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,768,3072,0.006650666809744305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,8192,0.053972442944844566
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,32,3072,0.005943111247486538
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,6144,8192,1.1812737782796223
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,768,8192,1536,0.020648888415760465
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,2560,256,0.00371111101574368
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,128,7168,0.04482311010360718
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,65536,256,0.041871110598246254
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,32,10240,0.013497778111033969
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,4096,8192,0.17734932899475098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,65536,128,0.01037599974208408
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,3584,3072,0.00925333301226298
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1536,5120,0.02379911144574483
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,512,8192,0.013384888569513956
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,5120,5120,0.019810666640599568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2560,6144,0.015440889530711703
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,7168,65536,5.624220530192058
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,6144,2560,0.01608088943693373
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,16384,1024,0.01759288873937395
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,768,12288,0.016855110724767048
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,64,6144,0.02238399949338701
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,8192,32,0.006147555592987273
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,256,65536,0.06693244642681546
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,1024,65536,0.06659822331534491
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,3072,2560,0.013264888690577613
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,16384,0.051543109946780734
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,32,0.006260444306664997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,768,512,0.004086222085687849
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,512,128,0.0064311110311084324
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,80,12288,2048,0.013569778038395775
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,3584,6144,0.013597333596812354
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,64,10240,64,0.006167111297448476
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8,6144,2560,0.011035555766688453
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,2048,2048,0.01257511145538754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,192,6144,3584,0.019326221611764696
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,32,10240,0.00618488879667388
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,32,1536,0.00619911112719112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,32,1536,0.004447111239035924
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,768,7168,0.06166488594479031
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,16384,2560,0.028512001037597656
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,2048,7168,0.015288889408111572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,6144,256,0.007326222128338284
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,256,7168,0.09884177976184422
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,384,16384,12288,0.216139554977417
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,3584,768,0.04010844561788771
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,3072,6144,0.012260444462299347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,6144,8192,0.028551999065611098
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,2560,2048,0.009172444542249044
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,16384,128,0.0035333335399627686
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,5120,0.022429333792792425
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,1024,1536,0.005673777725961473
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,10240,512,0.00565155554148886
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,128,2048,0.006286222073766921
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,12288,768,0.011100444528791638
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,2048,0.04537599947717455
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,80,8192,4096,0.023108444280094568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,6144,3072,0.040170666244294904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,2560,0.014669333895047506
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,7168,3072,0.012548444171746572
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,12288,5120,0.023267555567953322
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,768,3072,1536,0.012821333275900947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,768,128,0.0038702223036024305
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,192,2560,2560,0.009314667019579146
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4096,128,16384,0.041929778125551015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,12288,2560,0.015061333775520325
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,16384,6144,0.03621066610018412
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,2560,10240,0.020263110597928364
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,16384,10240,0.12100356154971653
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,1024,3584,0.007903999752468532
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,3072,64,0.006713777780532837
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,5120,64,0.003395555540919304
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,160,768,64,0.005476444545719359
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,16,12288,65536,0.41483910878499347
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,1536,16384,0.023934221929974023
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,512,7168,0.01107022249036365
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,32,3584,0.008406221866607666
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,16384,0.07180800040562947
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,128,65536,64,0.02369333306948344
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1536,1536,0.016350222958458793
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,10240,12288,0.07655022541681926
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,12288,7168,0.2638400130801731
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,512,2560,0.016588444511095684
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1,4096,768,0.0042088888585567474
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,160,2560,2560,0.008893333375453949
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8,3584,1024,0.005378666851255629
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,8192,6144,0.3036666711171468
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4096,256,16384,0.1228862206141154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,2048,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,16384,1024,0.021200888686709937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,7168,7168,0.027727999620967444
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,2560,5120,0.015403555499182807
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,5120,16384,0.03145600027508206
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,16,2048,3072,0.0073928890956772705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,768,8192,0.01536266671286689
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,48,2048,65536,0.07518933216730754
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,16384,1024,0.061689780818091504
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,128,65536,12288,0.22034666273328993
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,6144,12288,0.2081235514746772
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,10240,3072,0.01886399918132358
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,80,12288,128,0.0041946665280395085
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,7168,512,0.004540444248252445
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,4096,2048,0.008072888685597314
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,32,1024,0.008420444197124904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,64,1536,0.004512000001139111
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,16384,256,0.006607111129495833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,160,3584,64,0.0030417777597904205
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,2560,5120,0.05322133170233833
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,128,65536,256,0.015795555379655626
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,3584,8192,0.17321956157684326
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,32,3584,10240,0.0170160002178616
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,128,8192,0.013360000318951078
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,1,10240,2048,0.013731555806265937
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4096,10240,16384,1.0355653762817383
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,1024,16384,0.02953333324856228
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,12288,7168,0.10842933257420857
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,256,5120,65536,0.24877778689066568
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,96,10240,3584,0.02383733292420705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,4,3584,10240,0.022520888182852004
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,192,32,768,0.006474666711356904
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,10240,1024,0.020788444413079154
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,1,16384,12288,0.053605334626303784
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,256,5120,10240,0.037504888243145414
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2560,1536,0.006498666687144174
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,10240,3072,0.38446044921875
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,256,768,0.0047360000511010485
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,512,512,512,0.007303111255168915
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,96,16384,512,0.009005332986513773
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,48,4096,10240,0.019532443748580087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,4,3072,128,0.00397511116332478
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,10240,2560,0.014345778359307183
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,4,128,768,0.005514666852023866
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,7168,8192,0.08633599678675334
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,512,64,0.0026417778184016547
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,1024,6144,0.1286284393734402
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,32,128,3072,0.006508444332414203
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,64,5120,7168,0.015836444165971544
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,768,256,0.0034044444974925783
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,3584,1024,0.02144266664981842
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,1536,8192,0.01663377715481652
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,8192,256,12288,0.07705599731869169
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,96,10240,16384,0.0896186696158515
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,32,2048,6144,0.012627555264366997
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2048,6144,16384,0.3069529003567166
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,48,4096,4096,0.014555555250909595
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2,2560,1024,0.0063528890411059065
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8,256,32,0.005708444449636672
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,6144,12288,0.1142888863881429
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,2,3584,4096,0.01070844464831882
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,1024,1024,65536,0.18411289321051705
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,512,128,6144,0.008856000171767341
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,384,3584,2048,0.013089777694808112
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,256,7168,3072,0.01612355477280087
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2,2048,10240,0.01590666671593984
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,8192,1536,3584,0.10091822014914618
SGLang,0.5.9,NVIDIA H200,gemm,sglang,float16,2048,64,6144,0.012415111064910889
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8,16,128,7168,0.01092711091041565
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,6144,0.03633600142267015
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,512,2560,768,0.008014221986134848
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,10240,16384,0.21536710527208117
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,2048,1024,3584,0.023823110593689814
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,384,65536,3072,0.11785511175791423
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,8192,512,16384,0.23686488469441733
SGLang,0.5.9,NVIDIA H200,gemm,sglang,fp8_block,768,8192,3072,0.03506311111980014
