framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,16384,11.306951734754774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,12288,8.898372226291233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,16384,16.085247463650173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,12288,12.242124769422743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,12288,3.4272132449679904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,16384,5.4649768405490455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,10240,8.252806769476996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,12288,13.415537516276041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,10240,7.551823086208767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,10240,3.2599430084228516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,16384,17.58799574110243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,8192,6.975046369764541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,10240,10.13838619656033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,8192,5.859136793348525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,8192,2.5701963636610246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,7168,4.7219085693359375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,7168,2.249574237399631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,7168,7.435965643988715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,6144,4.32316631740994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,6144,5.369412316216363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,6144,2.056464936998155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,6144,5.006962670220269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,5120,5.838820563422309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,5120,3.5291786193847656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,5120,1.267489751180013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,4096,4.42051272922092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,4096,2.706208970811632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,4096,3.158820470174154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,4096,0.8127146826850044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,3584,4.624619377983941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,3584,2.298081715901693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,3584,2.816865709092882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,3584,0.9566720326741537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,3072,3.878790961371528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,3072,2.047779507107205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,3072,2.4369004567464194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,3072,0.8645128673977323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,2560,3.209405263264974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,2560,1.9793359968397353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,2560,2.1400898827446833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,2560,0.5133342213100857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,8192,6.654265933566624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,2048,1.3727519777086046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,2048,2.3368212381998696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,2048,0.5545475747850206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,2048,1.5765369203355577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,7168,5.884978824191624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,1536,1.8867244720458984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,1536,1.031707551744249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,1536,0.35637421078152126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,1536,1.1696418126424153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,1024,0.5532559818691677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,1024,0.7745173242357043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,1024,0.28314844767252606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,1024,0.915863143073188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,768,0.6659822463989258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,768,0.6024586889478896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,768,0.2759066687689887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,768,0.6731777720981174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,512,0.32984532250298393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,512,0.513797336154514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,512,0.28591820928785533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,512,0.5552657975090874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,256,0.2417128880818685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,256,0.4200026724073622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,256,0.4569270875718858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,128,0.19966044690873888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,128,0.43183112144470215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,128,0.199983106719123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,64,0.16465066538916692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,65536,256,0.20082933372921416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,65536,32,0.17049866252475313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,64,0.4532800250583225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,65536,32,0.47048086590237087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,5120,4.028112199571398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,65536,11.40089077419705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,65536,7.27590094672309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,16384,3.506257798936632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,65536,22.332522922092014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,16384,5.078914642333984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,16384,1.067829344007704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,16384,4.529296027289496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,12288,2.6647182040744357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,65536,19.976389567057293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,12288,3.257623036702474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,12288,0.8478186925252279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,10240,2.224301232231988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,12288,3.3032692803276906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,10240,2.349823209974501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,10240,0.7865111033121744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,10240,2.7716009351942272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,8192,2.2170018090142145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,8192,0.5128409067789713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,8192,1.795962651570638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,8192,2.0283946990966797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,7168,1.8575564490424261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,7168,1.4783893161349828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,7168,0.45879289839002824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,7168,1.6585893630981445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,6144,0.8864160113864474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,6144,1.2691431045532227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,65536,128,0.3217324415842692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,6144,1.3484622107611761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,6144,0.3217546674940321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,5120,0.9073040220472547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,5120,0.3908737765418158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,5120,1.0432106653849285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,4096,0.5467839770846897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,5120,1.1466506322224934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,4096,0.27721956041124135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,3584,0.523455089992947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,4096,0.8303048875596789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,4096,0.9172692828708224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,3584,0.19151022699144152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,3072,0.4064302179548476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,3584,0.7407129075792102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,3584,0.7910986476474338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,3072,0.16757244533962676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,3072,0.6243919796413845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,2560,0.37232266532050234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,3072,0.7037013371785482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,2560,0.5266995429992676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,2560,0.1524968942006429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,2048,0.29439467853969997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,2560,0.5973333252800835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,1536,0.21548799673716226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,2048,0.1724302238888211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,2048,0.4234773317972819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,2048,0.49096176359388566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,1024,0.15344799889458552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,1536,0.10263199938668145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,1536,0.33026666111416286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,1536,0.3891048961215549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,768,0.1156560050116645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,1024,0.24432712131076387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,1024,0.08303733004464044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,1024,0.37318221728007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,512,0.08976799911922878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,768,0.19193600283728707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,768,0.0752026637395223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,768,0.31342577934265137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,256,0.06552088922924466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,512,0.06736089123619927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,512,0.15578933556874594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,512,0.3243866761525472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,256,0.13281777169969347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,128,0.05387555559476217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,256,0.06070399946636624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,256,0.3371440039740668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,128,0.12170578373803033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,64,0.047255999512142606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,16384,128,0.0612915555636088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,16384,32,0.05009066727426317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,64,0.1247288915846083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,16384,32,0.12805155913035074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,65536,8.42840830485026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,65536,5.723755730523004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,65536,13.507657368977865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,16384,3.079843521118164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,65536,18.20367262098524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,16384,4.039213392469618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,16384,1.0313733418782551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,16384,3.3527128431532116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,12288,1.8874880472819011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,12288,0.496813350253635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,12288,2.8794017367892795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,10240,1.8119288550482855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,12288,2.4839555952284074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,10240,0.4751422140333388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,10240,1.8827137417263453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,8192,1.6343652937147353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,10240,2.1259483761257596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,8192,0.31143999099731445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,8192,1.4667867024739583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,7168,1.315768877665202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,8192,1.536087989807129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,7168,1.214472876654731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,7168,0.2925475438435872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,6144,0.8045528729756674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,7168,1.291376855638292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,6144,1.0164853201972115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,6144,0.23947644233703613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,5120,0.5324168735080296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,6144,1.1239502165052626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,5120,0.8655360009935168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,5120,0.20907555686102974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,4096,0.4167911211649577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,5120,0.9434053632948133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,4096,0.6891297764248319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,4096,0.16476089424557155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,3584,0.3764177693261041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,4096,0.7646817631191678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,3584,0.6326737933688694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,3584,0.14809155464172363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,3072,0.3155164453718397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,3584,0.6867164505852593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,3072,0.5110141966078017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,3072,0.13275021976894802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,2560,0.26491822136773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,3072,0.5946693420410156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,2560,0.4362355603112115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,2560,0.514276451534695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,2560,0.11791822645399307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,2048,0.21690666675567627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,2048,0.353441768222385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,2048,0.5830071237352159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,2048,0.09965422418382432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,1536,0.16724267270829943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,1536,0.2713235484229194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,1536,0.3370337751176622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,1536,0.08184178007973565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,1024,0.11668799983130561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,1024,0.19924533367156982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,1024,0.32481421364678276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,768,0.08919911252127753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,1024,0.06703822480307685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,768,0.1501013305452135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,512,0.0679253339767456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,768,0.3349413341946072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,768,0.06037866406970554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,512,0.12662222650316027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,512,0.054775999652014844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,256,0.05092089043723213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,256,0.1055333349439833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,256,0.33718487951490617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,256,0.05007288853327433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,128,0.09530222415924072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,512,0.28745510843065053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,12288,128,0.31726932525634766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,64,0.03929333223236932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,12288,128,0.0495760010348426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,32,0.04218399855825636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,64,0.09707733657624985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,12288,128,0.04185600082079569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,12288,32,0.10027377472983466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,65536,6.9239379035101996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,16384,128,0.3197084532843696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,65536,4.957064734564887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,16384,2.204338709513346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,65536,13.199927435980904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,16384,0.6478293206956651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,16384,3.4836701287163625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,65536,15.125230577256945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,16384,2.8630888197157116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,12288,0.4097626739078098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,12288,2.178269280327691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,12288,2.1163173251681857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,12288,2.1949564615885415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,10240,1.7828008863661025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,10240,0.3304639922247993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,10240,1.7409270604451497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,8192,0.9943040211995443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,10240,1.892093234592014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,8192,0.33312977684868705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,8192,1.2928222020467122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,7168,0.8112915356953939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,7168,1.0757697423299153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,7168,1.1775733100043404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,7168,0.259552001953125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,6144,0.6412151124742296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,8192,1.5154381857977972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,6144,0.24033244450887045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,6144,0.9143191443549262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,5120,0.4465795622931586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,6144,1.0174532996283638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,5120,0.1806417836083306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,4096,0.3665440082550049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,5120,0.7643751038445367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,5120,0.8531342082553439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,4096,0.1407182216644287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,4096,0.6159537633260092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,3584,0.335735109117296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,4096,0.7219911151462131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,3584,0.1684622234768338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,3072,0.2794160048166911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,3584,0.6212453312344021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,3072,0.11535199483235677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,3072,0.46378575430976016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,3072,0.6450373331705729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,2560,0.24485333760579428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,2560,0.10289244519339667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,2560,0.3876222239600287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,2560,0.46857865651448566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,2048,0.1955937809414334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,2048,0.3114782174428304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,2048,0.12303733825683594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,2048,0.385791990492079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,1536,0.13923022482130262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,1536,0.2419351206885444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,1536,0.07097155518001981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,1536,0.3153093390994602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,1024,0.0958168903986613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,3584,0.535080009036594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,1024,0.057322667704688184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,1024,0.17657777998182508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,1024,0.31789422035217285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,768,0.07628355423609416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,512,0.057413332992129855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,768,0.05350666575961643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,768,0.3446044392055935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,512,0.11148977279663086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,512,0.32029332054985893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,256,0.04223377837075127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,512,0.047157333956824414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,256,0.09223644600974189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,256,0.044237332211600415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,128,0.037109331952200994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,768,0.1393644412358602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,128,0.08231377601623535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,10240,128,0.04488355583614773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,64,0.030880000856187608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,64,0.08343022399478489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,10240,32,0.03239733311865065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,256,0.3939590983920627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,10240,32,0.08676355414920384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,65536,5.7416869269476996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,65536,4.060153961181641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,65536,10.010836283365885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,16384,2.0138320922851562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,65536,12.704259236653646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,16384,0.4334053463406033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,16384,2.564467536078559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,16384,2.998455895317925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,12288,1.019770622253418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,12288,0.30791110462612575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,12288,1.8071503109402125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,12288,1.8445839352077906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,10240,1.2537404166327584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,10240,1.393692440456814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,10240,1.5610435273912218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,10240,0.26212533315022785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,8192,0.5924533208211263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,8192,0.26573512289259166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,8192,1.126361740960015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,8192,1.219150225321452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,7168,0.6435093349880642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,7168,0.21919022666083443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,7168,0.9404711193508573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,6144,0.45568535063001847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,7168,1.057554668850369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,6144,0.1665697760052151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,6144,0.8016497823927137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,5120,0.3835288948482937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,6144,1.0240400102403429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,5120,0.6674666404724121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,5120,0.15035377608405218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,5120,0.7709146605597602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,4096,0.27812711397806805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,4096,0.11697155899471706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,4096,0.5356479750739204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,3584,0.26261067390441895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,4096,0.6308391359117296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,3584,0.1049377785788642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,3072,0.2205537822511461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,3584,0.7487520111931695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,3072,0.41365599632263184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,3072,0.4905698034498427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,3072,0.09371288617451985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,2560,0.17309510707855225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,3584,0.4694017834133572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,2560,0.08556444115108913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,2560,0.34275465541415745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,2560,0.43122845225863987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,2048,0.14452266693115234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,2048,0.07511022355821398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,2048,0.27429244253370494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,2048,0.3517279889848497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,1536,0.11383733484480117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,1536,0.06263822317123413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,1536,0.2108480003145006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,1536,0.32634221182929146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,1024,0.07901244693332247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,1024,0.15278311570485434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,1024,0.31385956870185006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,768,0.05855111281077067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,768,0.11943288644154866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,768,0.3469706641303168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,512,0.045040001471837364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,768,0.04526666800181071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,512,0.0962488916185167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,512,0.33930132124159074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,256,0.03547911180390252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,512,0.04081066780620151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,1024,0.05053066545062595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,256,0.07853688796361287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,256,0.03873422079616123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,128,0.02977777851952447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,256,0.32784443431430393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,128,0.06923466920852661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,8192,128,0.03807377815246582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,64,0.024442666106753882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,8192,32,0.025756445195939805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,64,0.06977599859237671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,8192,32,0.07215378019544813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,65536,5.155802832709418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,65536,9.9490966796875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,65536,3.29038577609592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,65536,9.06591542561849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,16384,1.7294719484117296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,16384,2.513765335083008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,16384,2.377424875895182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,16384,0.3822657797071669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,12288,1.1871795654296875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,12288,1.5464533699883356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,12288,1.7256640328301325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,12288,0.28255200386047363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,10240,128,0.28333067893981934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,10240,0.6128586663140191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,10240,0.4155680073632135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,10240,1.2447919845581055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,10240,1.4509706497192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,8192,0.6308613353305393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,8192,0.37102754910786945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,8192,1.0148809221055772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,7168,0.5595635308159722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,7168,0.8519946734110514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,7168,0.17356088426378039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,6144,0.357584900326199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,8192,1.1595777935451932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,6144,0.7335946824815539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,6144,0.8743484285142687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,7168,1.0092924965752494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,6144,0.1537084447013007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,5120,0.28962400224473744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,5120,0.13598399692111546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,5120,0.6163324250115289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,5120,0.7396302223205566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,4096,0.4982248942057292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,4096,0.6077502038743761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,4096,0.10556622346242268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,3584,0.20804533693525526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,4096,0.25506310992770725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,3584,0.4378257857428656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,3584,0.0972471104727851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,3584,0.5419626765780979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,3072,0.1835555500454373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,3072,0.08511555857128567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,3072,0.3769271108839247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,3072,0.4728257921006944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,2560,0.31706311967637807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,2560,0.4022986623975966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,2560,0.07649599843555026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,2048,0.13290399975246853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,2048,0.2543502118852404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,2048,0.33511998918321395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,1536,0.10013777679867214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,2560,0.15767289532555476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,2048,0.0654711127281189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,1536,0.05477422144677904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,1536,0.1950204504860772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,1536,0.3078106774224175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,1024,0.0707119968202379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,1024,0.04532000091340807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,1024,0.14078756173451742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,1024,0.3090453412797716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,768,0.05542400148179796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,768,0.1120871173010932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,512,0.041484445333480835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,768,0.3049386607276069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,512,0.08616355392667983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,512,0.30035556687249076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,256,0.030830221043692693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,512,0.03850222296184964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,256,0.06748000118467543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,256,0.3396204312642415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,256,0.035599999957614474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,128,0.026458667384253606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,768,0.0414008895556132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,128,0.05919466416041056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,7168,128,0.03536444571283128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,64,0.022267555197079975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,7168,128,0.3125102255079481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,7168,32,0.022686221533351477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,64,0.05779021978378296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,7168,32,0.05838933255937365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,65536,4.3501972622341585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,65536,2.422704060872396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,16384,1.9648017883300781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,65536,8.496572706434462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,65536,8.868030124240452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,16384,0.3317733340793186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,16384,2.082162645128038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,16384,2.1714053683810763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,12288,1.0067378150092232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,12288,0.48816442489624023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,12288,1.391576025221083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,10240,1.0257644653320312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,12288,1.6662836074829102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,10240,0.21158488591512045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,10240,1.134958267211914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,8192,128,0.31908178329467773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,8192,0.500998232099745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,8192,0.19585600164201525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,10240,1.383495118882921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,8192,1.1006133821275499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,7168,0.5399289131164551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,7168,0.26246221860249835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,7168,0.7851724624633789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,6144,0.4113919999864366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,7168,0.9610613716973199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,8192,0.9470248752170138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,6144,0.13487644990285239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,6144,0.6805609067281088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,5120,0.27814221382141113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,6144,0.8322835498385959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,5120,0.5697849061754015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,5120,0.12169778347015381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,4096,0.21693778038024902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,5120,0.701624870300293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,4096,0.09404444694519043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,4096,0.4619404474894206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,3584,0.1868844429651896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,3584,0.40426044993930393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,3072,0.15748266379038492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,3584,0.5082915623982748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,3584,0.08739644289016724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,3072,0.07746133539411756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,3072,0.35216087765163845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,3072,0.44592799080742734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,2560,0.13573066393534342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,4096,0.581767135196262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,2048,0.11518400245242649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,2560,0.29495467080010307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,2560,0.06866133213043213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,2560,0.3801271120707194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,1536,0.08501244253582424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,2048,0.05892355574501885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,2048,0.3398951159583197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,1536,0.1787182225121392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,1536,0.04829600122239855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,1536,0.3965146541595459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,1024,0.12813066111670599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,1024,0.041103111373053655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,1024,0.37146666314866805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,768,0.04799200097719828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,2048,0.2285146713256836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,768,0.10198133521609837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,512,0.033752000994152494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,768,0.3318897883097331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,768,0.038521776596705117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,1024,0.05845955345365736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,512,0.07860889037450154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,512,0.03558311197492812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,512,0.2821084393395318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,256,0.06176177660624186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,256,0.03291644321547614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,128,0.022833777798546687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,256,0.30348534054226345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,128,0.053469333383772105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,6144,128,0.030333333545260962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,64,0.019682667321628995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,32,0.020408888657887776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,64,0.05159999926884969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,6144,32,0.051890667941835195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,6144,256,0.03206933206982083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,6144,128,0.32795556386311847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,65536,1.747513771057129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,65536,3.9696197509765625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,16384,1.1560097800360787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,16384,1.7221430672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,16384,0.2838115427229139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,16384,2.010272979736328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,12288,0.8121431138780383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,65536,7.499104817708333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,12288,1.2515812979804146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,12288,0.2155484490924411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,10240,0.42644890149434406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,65536,7.842449612087673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,12288,1.5314284430609808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,10240,0.18530400594075522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,8192,0.3636808925204807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,10240,1.039571550157335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,8192,0.150163557794359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,10240,1.2825191285875108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,7168,0.3359831174214681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,8192,1.0365218056572807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,7168,0.13610577583312988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,7168,0.7265395588344998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,6144,0.2773706648084852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,7168,0.9003484514024523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,8192,0.8655164506700305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,6144,0.6285822126600478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,5120,0.22976533571879068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,6144,0.7821884685092502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,5120,0.524528874291314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,5120,0.6563742425706651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,4096,0.17250666353437635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,6144,0.12145333819919163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,4096,0.42483999994066024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,4096,0.5971662203470866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,4096,0.08435200320349799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,3584,0.1573759979671902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,3584,0.07665510972340901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,3584,0.37288710806104874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,3072,0.13740000459882948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,5120,0.10851289166344537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,3072,0.06894310977723864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,3072,0.3215120103624132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,2560,0.12340088685353597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,3072,0.4244071112738715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,2560,0.2692435582478841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,2560,0.39659733242458767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,2048,0.0993537770377265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,2560,0.060441778765784375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,3584,0.4818257755703396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,2048,0.215101334783766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,2048,0.33681689368353945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,1536,0.07652711206012301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,1536,0.16394578086005315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,1536,0.33815643522474503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,1024,0.05297333333227369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,1536,0.044129778941472374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,1024,0.11784977383083767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,1024,0.33372089597913956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,768,0.039628442790773176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,1024,0.03784977727466159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,2048,0.05122488737106323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,512,0.029003553920321997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,768,0.03410755594571432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,768,0.3306026723649767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,512,0.07193777958552043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,512,0.3271244366963704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,256,0.022309333086013794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,256,0.05490933193100823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,256,0.3095955583784315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,128,0.02018044392267863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,256,0.028946667909622192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,128,0.047019554509056934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,768,0.09384533431794907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,5120,128,0.2828142113155789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,64,0.01962488889694214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,5120,32,0.02027466727627648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,64,0.04487644301520454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,512,0.03138755427466498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,5120,32,0.04478844337993198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,5120,128,0.028198222319285076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,65536,1.5158497492472331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,65536,2.9290114508734812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,16384,1.0261191262139213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,16384,1.5361217922634547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,16384,0.24018488989935982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,12288,0.5997271007961696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,16384,1.8588648902045355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,65536,6.593736012776692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,12288,1.1318204667833116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,12288,0.18481866518656412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,10240,0.4231697718302409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,12288,1.4033626980251734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,65536,7.250614590115017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,8192,0.2814115683237712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,10240,1.1870044072469075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,10240,0.15829066435496011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,8192,0.7938568856981066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,7168,0.26338222291734487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,10240,0.9364737404717339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,8192,0.9630017810397677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,7168,0.11635555161370172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,7168,0.6710035536024305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,6144,0.2184542285071479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,7168,0.8504515753851997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,8192,0.12631821632385254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,6144,0.10523200035095215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,6144,0.5716524653964573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,5120,0.17706399493747285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,6144,0.7377813127305773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,5120,0.09187910954157512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,4096,0.14193422264522976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,5120,0.480479982164171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,5120,0.6233350965711806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,4096,0.07313422362009685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,4096,0.3877840042114258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,3584,0.1279253297381931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,4096,0.5136586825052897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,3072,0.11180088255140518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,3584,0.06935733556747437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,3584,0.3394977781507704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,3584,0.45406754811604816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,3072,0.06037244531843397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,2560,0.09246844715542263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,3072,0.29610665639241535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,3072,0.40592532687717015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,2048,0.07510666714774238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,2560,0.052248888545566134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,2560,0.24662489361233184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,2560,0.3347902297973633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,1536,0.058528886901007764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,2048,0.19611644744873047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,2048,0.04573955469661289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,2048,0.3266639974382189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,1024,0.04284711016549004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,1536,0.14887555440266928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,1536,0.0381315549214681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,1536,0.3333786593543158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,1024,0.10596711105770534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,768,0.03419555558098687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,1024,0.0321359998650021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,1024,0.33501333660549587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,512,0.024294222394625347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,768,0.08152444495095147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,768,0.02833777666091919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,768,0.3195626735687256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,512,0.06407733096016778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,256,0.01812000075976054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,512,0.025647999511824712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,512,0.3229120042588976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,256,0.04789244466357761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,128,0.014826666977670459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,256,0.34090932210286456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,128,0.04082311193148295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,4096,128,0.3045182228088379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,128,0.023058666123284235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,64,0.015272888872358533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,4096,32,0.016190222567982145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,64,0.0373964442147149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,4096,32,0.03663555449909634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,4096,256,0.0251920007997089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,65536,1.045837296379937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,65536,2.5061208936903214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,16384,0.938927968343099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,16384,1.4536195331149633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,16384,0.2151271104812622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,12288,0.3648666540781657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,16384,1.800138685438368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,12288,1.0743920008341472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,12288,0.1653075615564982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,65536,6.257512834337022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,12288,1.3605306413438585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,10240,0.32427554660373265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,10240,0.19799199369218615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,8192,0.23793244361877441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,10240,0.8947511249118381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,65536,6.993080139160156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,10240,1.1429519653320312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,8192,0.1177191072040134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,7168,0.22148356172773573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,8192,0.7436230977376302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,8192,0.9262950685289172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,7168,0.10540799962149726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,6144,0.18919911649492052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,7168,0.6330142021179199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,7168,0.8202533192104764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,6144,0.09648888640933567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,5120,0.1568062172995673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,6144,0.5432720184326172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,6144,0.7094604704115125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,4096,0.122143997086419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,5120,0.4586586422390408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,5120,0.08515022198359172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,5120,0.602216879526774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,3584,0.10861777597003514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,4096,0.36902578671773273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,4096,0.06531377633412679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,4096,0.48839733335706925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,3072,0.09620177745819092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,3584,0.0591813325881958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,3584,0.32408444086710614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,3072,0.3829413255055745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,3072,0.05484177668889364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,2560,0.08342844247817993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,2560,0.23478311962551543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,2560,0.04828800095452202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,3584,0.43829067548116046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,2048,0.06981244352128771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,3072,0.28004799948798287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,2048,0.1868871053059896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,2048,0.3346889019012451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,2048,0.0409804450141059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,1536,0.05127288897832235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,1536,0.140956441561381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,1536,0.034224887688954674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,2560,0.32712266180250377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,1024,0.03783911135461595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,1024,0.09917510880364312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,1024,0.027811557054519653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,768,0.02830577890078227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,1024,0.29779288503858775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,768,0.08020710945129395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,768,0.02513244416978624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,512,0.021568889419237774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,768,0.42074934641520184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,512,0.30367644627888996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,256,0.016647110382715862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,512,0.024646222591400146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,1536,0.2915884388817681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,256,0.044235554006364614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,128,0.014224888549910652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,256,0.41989866892496747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,128,0.03674666749106513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,128,0.020240889655219186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,64,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3584,128,0.4286506705813938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3584,32,0.012551111479600271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,64,0.03412266572316488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3584,256,0.021372444099850122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,32,0.03353244397375319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3584,512,0.06005866660012139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,65536,2.3775511847601996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,65536,1.161757363213433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,16384,0.4943688710530599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,16384,1.3665581809149847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,16384,1.749544037712945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,16384,0.2001404497358534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,12288,0.3451440069410536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,65536,5.7896677652994795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,12288,0.15137688318888345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,10240,0.25285156567891437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,65536,6.820606231689453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,12288,1.3278862635294597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,12288,1.0133448706732857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,10240,0.13474399513668484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,10240,0.8478924433390299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,8192,0.21075110965304902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,8192,0.10866755247116089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,10240,1.1164933310614693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,7168,0.18473154968685576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,8192,0.6991830931769477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,8192,0.9049243927001953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,7168,0.09816355837715997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,6144,0.1559413274129232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,7168,0.5983155568440756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,7168,0.8039609061347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,6144,0.519843578338623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,6144,0.0888026687833998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,5120,0.1469635566075643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,6144,0.6920995712280273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,5120,0.0799937778049045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,4096,0.1042319999800788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,5120,0.4339866638183594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,5120,0.5877928733825684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,4096,0.061753776338365346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,4096,0.39118843608432347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,3584,0.10214310884475708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,4096,0.4814675649007161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,3072,0.08764800098207261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,3584,0.05510044429037306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,3584,0.3089955647786458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,3584,0.42756978670756024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,2560,0.07867288589477539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,3072,0.2651786539289686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,3072,0.43056355582343203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,2560,0.22046311696370444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,2560,0.3378257751464844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,2048,0.060025778081681996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,2560,0.043064001533720225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,2048,0.1710275544060601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,2048,0.03706489006678263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,2048,0.3418097760942247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,3072,0.04942489001486036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,1536,0.13133955001831055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,1536,0.33085510465833873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,1024,0.03233688738611009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,1536,0.030368887715869482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,1024,0.0943031112353007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,768,0.026525333523750305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,1024,0.024539555112520855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,1024,0.44851287206013996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,768,0.0752026637395223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,512,0.019998222589492798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,768,0.3300106790330675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,1536,0.0481680002477434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,768,0.023079999619060095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,512,0.05626489056481255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,256,0.014699555105633207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,512,0.020840888222058613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,512,0.32756445142957896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,256,0.019312888383865356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,256,0.04042044281959534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,128,0.033632887734307185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,256,0.3889591164059109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,64,0.0118231111102634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,3072,128,0.30591734250386554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,3072,128,0.018876444962289598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,32,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,64,0.030795554320017498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,3072,32,0.030113776524861652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,3072,128,0.013507555756303998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,65536,0.7386017905341254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,65536,2.1190569135877824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,16384,0.5615875456068251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,16384,1.2762319776746962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,16384,0.18058399359385172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,12288,0.29998665385776097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,16384,1.680992020501031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,12288,0.9581946267022027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,12288,0.13971377743615046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,12288,1.275952868991428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,65536,5.354593912760417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,10240,0.261757320827908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,10240,0.1231457789738973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,8192,0.20784444279140898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,10240,0.8045520252651639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,10240,1.0741999944051106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,65536,6.529215918646918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,8192,0.6515208880106608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,8192,0.09721244706047906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,7168,0.17010045051574707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,8192,0.8693146705627441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,7168,0.08752177821265326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,6144,0.15650755829281277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,7168,0.5644417868720161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,7168,0.7697946760389539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,6144,0.48788801829020184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,5120,0.13158399528927273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,6144,0.6706177923414441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,6144,0.08029066854053073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,5120,0.07194933626386854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,5120,0.4093422094980876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,5120,0.5679501957363552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,4096,0.10273155238893296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,4096,0.05484799875153435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,4096,0.32986222373114693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,3584,0.0914906660715739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,4096,0.4667004479302301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,3584,0.04991555545065138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,3584,0.290047115749783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,3072,0.08212977647781372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,3584,0.4553235371907552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,3072,0.251601775487264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,2560,0.06828800174925062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,3072,0.36293334431118435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,3072,0.04463288850254483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,2048,0.056236445903778076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,2560,0.038324445486068726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,2560,0.20948889520433214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,2048,0.03186488813824124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,2560,0.46354044808281797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,1536,0.04348799917432997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,2048,0.339030212826199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,2048,0.1662328905529446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,1024,0.030125333203209773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,1536,0.027506666051016912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,1536,0.12337688604990642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,1536,0.3281493451860216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,768,0.02513422237502204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,1024,0.0885537796550327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,1024,0.021760889225535925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,1024,0.3211119969685872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,512,0.01819022165404426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,768,0.07015644179450141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,768,0.020085333122147452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,512,0.05091555582152473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,256,0.013395555317401886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,768,0.39678578906589085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,512,0.01891999940077464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,512,0.2758115662468804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,128,0.01072088877360026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,256,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,256,0.036738667223188616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,128,0.03030844529469808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,64,0.009655999640623728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2560,128,0.01756888959142897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,256,0.41096266110738117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2560,32,0.009694221946928237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2560,128,0.333858675426907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,64,0.027464888162083093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2560,32,0.027237332529491846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,65536,0.609133349524604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,65536,1.4907555050320094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,16384,0.3204302257961697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,16384,1.2078577677408855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,16384,0.15729955832163492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,12288,0.23049778408474395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,16384,1.6118933359781902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,12288,0.9076862335205078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,12288,0.12129067050086127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,12288,1.223139550950792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,10240,0.20313421885172525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,65536,4.921786838107638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,10240,0.10974044269985622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,10240,0.7584186659918891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,8192,0.15903822580973306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,10240,1.0303067101372612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,8192,0.08516089121500652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,8192,0.6151457892523872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,7168,0.13743644290500218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,8192,0.8362017737494575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,7168,0.07836355765660603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,6144,0.11921421686808269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,7168,0.7376764615376791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,6144,0.4586168924967448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,65536,6.255686018202041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,6144,0.07162400086720784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,6144,0.6387831370035807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,5120,0.09839999675750732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,7168,0.5341208775838217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,5120,0.06286844280030993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,5120,0.38508089383443195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,4096,0.07946222358279757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,5120,0.541412459479438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,3584,0.07489955425262451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,4096,0.31164444817437065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,4096,0.04814844330151876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,4096,0.4441591103871663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,3072,0.06252444452709623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,3584,0.0432533323764801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,3584,0.3933093282911513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,3072,0.23717244466145834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,3072,0.03793777690993415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,2560,0.050075554185443454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,3072,0.45517598258124453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,2560,0.19762755764855278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,2560,0.3790026770697699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,2560,0.033442666133244835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,2048,0.040892445378833346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,3584,0.27587466769748265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,2048,0.155184891488817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,2048,0.027463111612531874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,2048,0.39243555068969727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,1536,0.11567733022901748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,1536,0.4139404561784532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,1024,0.02314844396379259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,1536,0.022408889399634466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,1024,0.08042844136555989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,1024,0.3788284460703532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,1024,0.019609777463806998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,768,0.018054222067197163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,1536,0.03162844313515557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,768,0.06548444430033366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,768,0.01812977757718828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,512,0.013889777991506787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,768,0.3267760011884901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,256,0.010873777998818291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,512,0.047692444589402944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,256,0.03352888756328159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,512,0.01735466718673706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,512,0.33328355683220756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,128,0.009744889206356471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,256,0.016176889340082806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,256,0.33338401052686906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,128,0.027095110880004034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,64,0.009319999979601966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,2048,128,0.016209777858522203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,64,0.024694222542974684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,2048,32,0.024344889654053584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,2048,128,0.5168177816602919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,65536,0.9322409100002713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,2048,32,0.009672889278994666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,16384,0.22593066427442762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,65536,0.6268977589077419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,16384,1.129991955227322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,16384,0.1396933396657308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,12288,0.17515822251637778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,16384,1.5765377680460613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,12288,0.8507342338562012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,12288,0.10749244689941406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,10240,0.15192355049981013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,65536,4.55861070421007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,12288,1.1967279646131728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,10240,0.0985200007756551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,8192,0.12018399768405491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,10240,0.7105058034261068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,10240,1.0066204071044922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,8192,0.5725688934326172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,7168,0.1014675564236111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,8192,0.07645333475536771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,8192,0.8143875333997938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,65536,6.126509348551433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,7168,0.06901244322458903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,6144,0.0905039972729153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,7168,0.5007946756150987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,7168,0.7211857901679145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,6144,0.06449511316087511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,5120,0.07565866576300727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,6144,0.43252889315287274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,6144,0.6240817705790201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,5120,0.057146668434143066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,5120,0.36273601320054794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,4096,0.06586488750245836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,5120,0.5319822099473741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,3584,0.05441955394215054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,4096,0.043360888957977295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,4096,0.4347413380940755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,3584,0.25924799177381724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,3072,0.04760266674889458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,3584,0.4629511303371853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,3072,0.2235679891374376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,3072,0.03377244538731045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,2560,0.04043644335534837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,3072,0.45169512430826825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,4096,0.29250576761033803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,2560,0.1872746679517958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,2560,0.02903111113442315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,2048,0.03290844294759963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,3584,0.039312889178593956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,2560,0.4328888787163629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,1536,0.027089777919981215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,2048,0.14664355913798013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,2048,0.02398577829202016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,2048,0.32351022296481663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,1024,0.01907822158601549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,1536,0.020633776982625324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,1536,0.10838488737742107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,1536,0.3209022151099311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,1024,0.07427466577953763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,768,0.015657777587572735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,1024,0.3253546555836995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,1024,0.0178195552693473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,768,0.05686933464474148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,512,0.013052444491121503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,768,0.319888883166843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,768,0.016879111528396606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,512,0.0441599984963735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,256,0.010040889183680216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,512,0.01588000026014116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,256,0.029802666770087347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,512,0.45187822977701825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,128,0.00907466643386417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,256,0.015150222513410779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,256,0.37723377015855575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1536,128,0.014920888675583733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,64,0.009336888790130615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1536,128,0.3452960120307074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1536,32,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,64,0.02199911077817281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,32,0.021309332715140447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,65536,0.6556631194220649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1536,128,0.024039111203617517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,16384,0.15048000547620985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,65536,0.4497217602199978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,16384,1.058150185479058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,16384,0.11880621645185684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,12288,0.11860533555348714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,16384,1.5045724444919164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,12288,0.7963653140597873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,12288,1.1423217985365126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,12288,0.09119021892547607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,65536,4.217859480116102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,10240,0.0990755558013916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,10240,0.08262666728761461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,8192,0.08067199918958876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,10240,0.6652035713195801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,10240,0.9627350701226128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,8192,0.5372675789727105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,8192,0.06452888912624784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,7168,0.07076088587443034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,8192,0.7770062022738986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,7168,0.058766219351026744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,7168,0.47125864028930664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,6144,0.06204533576965332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,7168,0.687736882103814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,65536,5.8537953694661455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,6144,0.4055795669555664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,5120,0.05515733361244202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,6144,0.05465333329306709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,6144,0.593140443166097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,4096,0.04397955536842346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,5120,0.04997155401441786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,5120,0.3409742249382867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,5120,0.5064000023735894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,4096,0.037263999382654824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,4096,0.27000000741746694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,4096,0.4375475512610541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,3584,0.2424000104268392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,3584,0.36804623074001735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,3072,0.03221777743763394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,3584,0.03253066539764404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,3072,0.028353777196672227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,3072,0.3422257900238037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,2560,0.02796799937884013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,2560,0.1747386720445421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,3584,0.0406275557147132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,2048,0.023344000180562336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,2560,0.0244862222009235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,2048,0.13704710536532932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,2048,0.3330649005042182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,2048,0.02087644404835171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,1536,0.018224888377719455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,3072,0.2078933318456014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,1536,0.10089066955778335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,1024,0.013735999663670858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,1536,0.018019555343521964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,1536,0.3306737740834554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,1024,0.06919999917348225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,1024,0.015970667203267414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,768,0.011362666885058085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,1024,0.3125724527570936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,2560,0.4311102231343587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,512,0.008861333131790161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,768,0.052404443422953285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,768,0.014856888188256158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,768,0.3262977865007189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,256,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,512,0.014182221558358936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,512,0.03778666588995192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,512,0.3344702190823025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,128,0.006311999840868845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,256,0.026670222481091816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,256,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,256,0.31711112128363717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,128,0.021306667062971327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,64,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,1024,32,0.005984000033802456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,1024,128,0.013807111316257052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,1024,128,0.29983199967278373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,64,0.018958222534921434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,1024,32,0.018195556269751657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,65536,0.4627031220330133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,16384,0.1429448922475179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,65536,0.43904709815979004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,16384,1.0240275065104167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,16384,1.4680453406439886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,16384,0.11637777752346462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,12288,0.090865777598487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,12288,0.7703742451137967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,65536,4.056342230902778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,10240,0.08155822091632418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,12288,0.09078400002585517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,12288,1.1144097646077473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,10240,0.0805164443122016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,8192,0.0666746695836385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,10240,0.6446720229254829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,10240,0.9405724207560221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,8192,0.06365778048833211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,65536,5.721660614013672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,7168,0.060933331648508705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,8192,0.5192560089959039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,7168,0.05819911426968045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,8192,0.7634995778401693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,6144,0.05158311128616333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,7168,0.45627111858791775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,7168,0.6734746827019585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,6144,0.05340711275736491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,5120,0.043509334325790405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,6144,0.39295199182298446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,6144,0.581376870473226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,4096,0.034832000732421875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,5120,0.3299653265211317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,5120,0.048844443427191846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,5120,0.4951137966579861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,3584,0.031268444326188825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,4096,0.2664773464202881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,4096,0.4058817757500543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,3584,0.2341609001159668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,3072,0.02697155541843838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,3584,0.031073778867721558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,3584,0.3614799976348877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,3072,0.027313777142100867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,3072,0.20251022444831002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,2560,0.02277155551645491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,3072,0.33791467878553605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,2560,0.02425066630045573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,2560,0.16957599586910674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,2048,0.01922044489118788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,2560,0.3409324487050374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,4096,0.03604888916015625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,1536,0.014735110931926303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,2048,0.020595555504163105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,2048,0.1326835552851359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,2048,0.32426489724053276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,1024,0.011584888729784222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,1536,0.0181031111213896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,1536,0.09738044606314765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,1536,0.3157564534081353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,768,0.009186666872766282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,1024,0.06810844606823392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,1024,0.016128889388508268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,1024,0.3325057824452718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,768,0.051517334249284535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,512,0.007407111426194509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,768,0.329382234149509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,768,0.015142222245534262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,512,0.036158220635520086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,256,0.006695999867386288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,512,0.014188443620999655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,512,0.3221653302510579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,128,0.006246222390068903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,256,0.025390222668647766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,256,0.013713777893119387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,128,0.019575999842749704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,64,0.006096000058783426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,256,0.4413137700822618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,768,128,0.01346577786737018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,768,128,0.29997067981296116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,768,32,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,64,0.01754666699303521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,768,32,0.016843555702103507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,65536,0.31877867380777997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,16384,0.08213866419262357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,65536,0.3553617795308431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,16384,0.9862186643812391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,16384,0.09588533639907837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,12288,0.06100800302293566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,16384,1.4347235361735027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,12288,0.7424675623575846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,12288,0.07376622491412692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,10240,0.0548622210820516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,12288,1.0883590910169814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,10240,0.6207288636101617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,65536,3.9165848626030813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,8192,0.043975111511018544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,10240,0.06627288791868421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,10240,0.9154195785522461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,8192,0.052342222796546094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,7168,0.0402782228257921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,8192,0.5001129044426812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,8192,0.7412346733940972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,7168,0.43924887975056964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,7168,0.048178666167789035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,6144,0.03429777754677667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,7168,0.6531039873758951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,65536,5.58038330078125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,6144,0.3775155544281006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,5120,0.028656889994939167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,6144,0.5702177683512369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,6144,0.04519911275969612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,5120,0.3174613316853841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,5120,0.041161778900358416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,4096,0.023895111348893907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,5120,0.48177242279052734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,4096,0.02906755606333415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,4096,0.25587111049228245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,3584,0.021643555826610986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,4096,0.40937333636813694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,3584,0.02605777813328637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,3584,0.2254719999101427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,3072,0.017875555488798354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,3584,0.38344621658325195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,3072,0.02239911092652215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,3072,0.19507111443413627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,2560,0.016172443826993305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,3072,0.38581511709425187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,2560,0.020261334048377145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,2560,0.16264443927341038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,2048,0.014139556222491793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,2560,0.3455306688944499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,1536,0.01129333343770769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,2048,0.12725244628058538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,2048,0.334262212117513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,1536,0.09369422329796685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,2048,0.017431111799346078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,1024,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,1536,0.01515111161602868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,1536,0.32101689444647896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,768,0.007689778175618913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,1024,0.013770666387346057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,1024,0.33128354284498424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,768,0.048807998498280845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,512,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,768,0.01313777764638265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,768,0.324947542614407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,512,0.034536888202031456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,256,0.0052284445199701525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,512,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,512,0.3383359909057617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,256,0.023726221587922838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,128,0.004592000196377437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,256,0.01202933324707879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,256,0.32304265764024526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,128,0.01792266633775499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,64,0.004881777697139316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,512,128,0.011874666644467248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,1024,0.06502399841944377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,512,32,0.005239999956554837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,64,0.01588888963063558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,65536,0.20699022875891793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,512,32,0.015603555573357476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,65536,0.35228978263007266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,16384,0.056681778695848256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,512,128,0.31697599093119305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,16384,0.9554853439331055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,16384,0.09585599766837226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,12288,0.04360266526540121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,12288,0.7203066613939074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,16384,1.3984781901041667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,65536,3.774383121066623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,10240,0.03705599904060364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,12288,0.07369422250323825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,12288,1.0624141693115234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,10240,0.6008257865905762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,8192,0.029719110992219713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,10240,0.06678933567470975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,10240,0.8930382198757596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,8192,0.052079998784595065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,7168,0.027496889233589172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,65536,5.434698740641276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,8192,0.7242693371242948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,7168,0.4253519905938043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,6144,0.02160088883505927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,7168,0.04862222075462341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,7168,0.6398293177286783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,6144,0.36653955777486164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,5120,0.015804444750150044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,6144,0.04492177897029453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,8192,0.484374205271403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,6144,0.5539297527737087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,4096,0.013361777696344586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,5120,0.0414737794134352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,5120,0.30770932303534615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,5120,0.46866665946112734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,3584,0.01239111108912362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,4096,0.02825333343611823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,4096,0.3874888949924045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,4096,0.24907377031114367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,3072,0.010880000061459012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,3584,0.026145777768558923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,3584,0.21806932820214164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,3584,0.34464266565110946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,2560,0.010100444157918295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,3072,0.022796443767017786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,3072,0.18886489338345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,3072,0.31255022684733075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,2560,0.15723378128475612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,2048,0.008344888687133789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,2560,0.020199111766285367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,2560,0.36461512247721356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,2048,0.1234017743004693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,2048,0.017606221967273288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,2048,0.3702204492357042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,1536,0.09018755621380276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,1536,0.01516355574131012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,1536,0.38350311915079754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,1024,0.006079111248254776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,1024,0.06173421939214071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,1024,0.38077955775790745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,768,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,1536,0.007341333561473423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,768,0.04536622100406223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,768,0.35781155692206484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,768,0.01313333296113544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,512,0.004842666702138053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,512,0.0328435566690233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,256,0.0042275554604000514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,512,0.3439146677652995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,512,0.012869333227475485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,256,0.022093334131770666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,128,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,1024,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,256,0.012159999873903064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,256,0.3243955506218804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,64,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,256,128,0.2958391242557102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,64,0.014525334040323893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,256,32,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,256,128,0.011962666279739805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,32,0.014112000664075216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,65536,0.16321688228183323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,65536,0.36884623103671604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,256,128,0.016528000434239704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,16384,0.04317155480384827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,16384,0.9379670884874133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,16384,0.09834933280944824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,12288,0.034627556800842285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,12288,0.7080080244276258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,65536,3.7056736416286893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,12288,0.07744622230529785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,10240,0.03018666638268365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,10240,0.5907831192016602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,16384,1.3932658301459417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,10240,0.06850933366351657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,8192,0.024522667129834492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,8192,0.47553422715928817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,8192,0.724213335249159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,8192,0.05499733487764994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,7168,0.02160533269246419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,7168,0.41692267523871523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,7168,0.6372257868448893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,7168,0.050203555160098605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,6144,0.014108444253603617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,12288,1.0563288794623482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,6144,0.35911112361484104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,6144,0.045888000064426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,5120,0.011113777756690979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,6144,0.5526133113437229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,5120,0.30138222376505536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,5120,0.4694364335801866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,5120,0.04269688990381029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,4096,0.010310221877362993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,4096,0.24389423264397514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,4096,0.38628623220655656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,3584,0.008735111190213097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,4096,0.02975644336806403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,3584,0.21473066012064615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,65536,5.423301272922092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,3584,0.34756000836690265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,3584,0.026299556096394856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,3072,0.007832888927724626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,3072,0.02366222275627984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,2560,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,3072,0.3485555648803711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,2560,0.1546808878580729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,2560,0.3114559915330675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,2048,0.006502222269773483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,2560,0.020940444535679285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,2048,0.12045599354637994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,2048,0.3101591004265679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,1536,0.005675555517276128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,2048,0.01795022189617157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,10240,0.8898382186889648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,3072,0.18477066357930502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,1024,0.004889777965015835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,1536,0.08823466963238186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,1536,0.015704888436529372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,1536,0.3249217669169108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,1024,0.06069778071509468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,768,0.004566222015354368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,1024,0.014095111025704278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,1024,0.31675555970933705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,512,0.0041493334703975255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,768,0.3216959900326199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,768,0.013524444566832649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,512,0.03193599979082743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,512,0.012758221891191272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,256,0.003569777641031477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,256,0.021514667405022517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,768,0.044553776582082115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,256,0.3897617657979329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,256,0.011878222227096558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,512,0.3741155465443929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,128,0.01588000026014116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8192,128,128,0.011762667033407422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,64,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,64,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,32,0.0032551110618644287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,128,128,0.0036915557252036203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,128,32,0.013501333693663279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,16384,0.047372443808449634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,65536,0.1578622261683146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,12288,0.034437331888410784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,16384,0.9287644492255317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,10240,0.029157333903842505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,12288,0.7009706497192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,8192,0.0237137774626414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,10240,0.5845964219835069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,7168,0.021273778544531927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,8192,0.47193866305881077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,6144,0.01293955577744378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,7168,0.41262488894992405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,5120,0.009745777481132084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,6144,0.35591912269592285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,4096,0.008750222623348236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,5120,0.29822222391764325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,3584,0.007909333540333642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,4096,0.2421893278757731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,65536,3.6706089443630643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,3072,0.006993778049945831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,3584,0.21238577365875244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,2560,0.007124444676770105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,2048,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,3072,0.18287732866075304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,2560,0.15164889229668513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,1536,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,1024,0.004549333204825719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,2048,0.11906399991777207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,1536,0.08681155575646295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,768,0.0046657779150538975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,512,0.003918222255176968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,1024,0.05931555562549167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,256,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,768,0.043897777795791626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,128,0.0032133333798911837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,256,0.020963556236690942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,64,0.0032435556252797446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,128,0.01521511044767168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,64,32,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,64,0.013439999686347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,32,0.013120888835854001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,65536,0.1579351160261366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,16384,0.04788622260093689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,64,512,0.03156800071398417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,12288,0.031982223192850746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,16384,0.9292462666829427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,10240,0.027541332774692114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,12288,0.7000888718499078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,8192,0.02267733381854163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,10240,0.5835653411017524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,7168,0.02030044462945726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,8192,0.4707057740953233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,6144,0.011199110911952125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,7168,0.4120186699761285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,5120,0.009359999663299983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,6144,0.35572089089287656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,5120,0.2974693245357937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,4096,0.007697777615653143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,3584,0.007323555648326874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,4096,0.2418524424235026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,3072,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,65536,3.6678309970431857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,3584,0.21428888373904756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,2560,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,2048,0.005222222043408289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,3072,0.18563289112514922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,1536,0.005582222094138463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,2560,0.15427377488878039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,2048,0.12062044938405354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,1024,0.004581333448489507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,768,0.004264889078007804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,1536,0.08825421995586819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,512,0.003887999802827835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,1024,0.059798220793406166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,256,0.0035386664999855887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,768,0.04404177930619982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,512,0.03166844447453817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,128,0.0032319999817344877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,64,0.002955555501911375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,256,0.020920000142521326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8192,32,32,0.0029031110720502008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,128,0.015126221709781222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,32,0.01276088919904497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8192,32,64,0.01347733371787601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,16384,5.632724338107639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,16384,6.211643642849392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,16384,2.4478123982747397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,16384,6.664160834418403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,12288,4.22831302218967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,12288,4.365805308024089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,12288,1.5719724231296117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,12288,4.732615152994792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,10240,3.9458953009711375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,10240,3.4774888356526694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,10240,1.203287124633789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,10240,3.906904856363932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,8192,3.495653364393446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8192,128,128,0.3025040096706814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,8192,0.9018497467041016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,8192,3.0291307237413196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,8192,3.048938751220703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,7168,2.906038284301758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,7168,0.6528328789605035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,7168,2.304541269938151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,7168,2.6001466115315757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,6144,2.0051146613226996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,6144,2.4274879031711154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,6144,0.5444569057888454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,6144,2.3757466210259333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,5120,1.8232106102837458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,5120,0.6430008676317003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,5120,1.6432061725192602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,5120,1.8113572862413194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,4096,1.337646272447374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,4096,1.4225876066419814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,4096,0.45278135935465497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,4096,1.6596817440456813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,3584,1.1840231153700087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,3584,0.38115466965569395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,3584,1.2010782029893663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,3584,1.2451777988009982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,3072,1.1346196068657768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,3072,1.0269226498074002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,3072,0.3558515442742242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,3072,1.0832915835910373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,2560,0.731601768069797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,2560,0.3451537821027968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,2560,0.8239537874857584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,2560,0.8236159748501248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,2048,0.8029386732313367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,2048,0.7903315756056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,2048,0.2271733283996582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,2048,0.7132382392883301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,1536,0.585800912645128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,1536,0.1872791051864624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,1024,0.35747289657592773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,1536,0.5560026698642307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,1536,0.5615022447374132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,1024,0.15509155061509874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,768,0.23542043897840711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,1024,0.40098932054307723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,1024,0.4328266779581706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,768,0.1374595562616984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,768,0.3163111209869385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,512,0.1747431092792087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,768,0.41508089171515566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,512,0.12362755669487847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,512,0.26673955387539333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,512,0.3448364469740126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,256,0.10744711425569321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,256,0.23651554849412706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,256,0.3474213282267253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,128,0.10485244459576076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,128,0.2247920036315918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,65536,128,0.3172222243414985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,64,0.08535555336210464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,32,0.08824533224105835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,64,0.2308879958258735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,65536,256,0.12693244881100124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,65536,32,0.23941334088643393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,65536,128,0.10573955376942952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,65536,2.7944293551974826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,16384,1.376620398627387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,65536,6.27559831407335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,16384,1.8141074710422094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,16384,0.3708631197611491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,16384,2.3244516584608292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,65536,9.144465128580729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,12288,1.6875342263115778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,65536,8.132155524359808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,12288,0.28088445133633083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,12288,1.3254559834798176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,12288,1.3607147004869249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,10240,1.2460826237996419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,10240,1.0702826182047527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,10240,0.23780534002516004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,10240,1.1071884367201064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,8192,0.1940302186542087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,8192,1.1736586888631184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,8192,0.9015910890367297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,7168,0.6291759808858236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,8192,0.8846142027113172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,7168,0.19273689058091906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,6144,0.54602935579088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,7168,0.7383564313252767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,7168,0.7706471019321017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,6144,0.15743288728925917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,6144,0.6478586726718479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,6144,0.6689715385437012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,5120,0.3559262222713894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,5120,0.5268399980333116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,5120,0.15939199924468994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,4096,0.28655733002556694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,5120,0.5788942442999946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,4096,0.41420711411370176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,4096,0.10764888922373454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,3584,0.2552559905582004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,4096,0.48405689663357204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,3584,0.09753511349360149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,3584,0.36348621050516766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,3584,0.42871822251213926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,3072,0.22660888565911186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,3072,0.08983111381530762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,3072,0.3138186666700575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,3072,0.37593687905205625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,2560,0.07771644327375624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,2560,0.2681573232014974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,2048,0.15306222438812256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,2560,0.45418490303887266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,2048,0.21994400024414062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,2048,0.0660604437192281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,1536,0.10759911272260879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,2560,0.1716924508412679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,1536,0.17080088456471762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,1536,0.35399821069505477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,1024,0.07294133636686537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,1536,0.05491822295718723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,1024,0.12869600454966226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,768,0.06084622277153862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,1024,0.046388443973329335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,1024,0.4360728793674045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,768,0.10226666927337646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,2048,0.33113156424628365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,512,0.04679200053215027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,768,0.04307466745376587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,768,0.39848799175686306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,512,0.08453955252965291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,256,0.034345779154035784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,512,0.33018043306138783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,512,0.03945866558286879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,128,0.029504001140594482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,256,0.07229422198401557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,256,0.03345155715942383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,256,0.3646559980180528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,64,0.024476443727811176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,128,0.06791110833485921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,16384,128,0.034127109580569796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,16384,128,0.3042088879479302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,16384,32,0.025177778469191656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,64,0.06880088647206624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,16384,32,0.07027022043863933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,65536,2.335040834214952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,16384,1.5495306650797527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,65536,4.53373760647244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,16384,1.4838746388753254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,16384,1.5839964548746746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,16384,0.29010311762491864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,12288,0.9061928855048286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,65536,7.146745469835069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,12288,1.0728498035007052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,65536,6.275863223605686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,12288,0.22137422031826445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,10240,0.6485297944810655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,10240,0.8821502261691623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,10240,0.2868186632792155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,12288,1.114381366305881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,10240,0.9275698131985135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,8192,0.15238400300343832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,8192,0.48886754777696395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,7168,0.3862924575805664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,8192,0.7428320248921713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,8192,0.7607857916090223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,7168,0.1373360024558173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,7168,0.6296035448710123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,6144,0.33718400531344944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,6144,0.12303911315070258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,7168,0.6702444288465711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,5120,0.27830754386054146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,6144,0.5793164571126302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,5120,0.1048168871137831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,5120,0.4322648843129476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,4096,0.2168275515238444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,5120,0.5825662083095974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,4096,0.08917333020104302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,4096,0.3422524399227566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,6144,0.523255983988444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,4096,0.4064479933844672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,3584,0.07469689183764987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,3584,0.29756177796257866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,3072,0.15633066495259604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,3072,0.25586220953199595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,3072,0.35293955273098415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,2560,0.1306355529361301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,3584,0.18854133288065592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,3072,0.0689893364906311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,2560,0.2198408842086792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,3584,0.36983288658989805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,2560,0.3797004487779405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,2560,0.06037955813937717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,2048,0.10771644115447998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,2048,0.05270488725768196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,1536,0.08965955840216742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,2048,0.42319644822014707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,1536,0.14172444078657362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,1536,0.3870844576093886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,1536,0.04453511039415995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,1024,0.05871466795603434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,1024,0.10539911190668742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,2048,0.18094933032989502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,1024,0.38112176789177793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,768,0.04641244477695889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,1024,0.03794577717781067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,768,0.08433866500854492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,768,0.03460799985461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,512,0.035207112630208336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,768,0.34872889518737793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,512,0.07006488906012641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,256,0.03225866622394986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,512,0.03198488884501987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,512,0.3205164538489448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,128,0.02276533345381419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,256,0.05889066722657946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,256,0.028788444068696763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,128,0.05464888943566216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,256,0.38234488169352215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,64,0.01963733302222358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,12288,128,0.02805866797765096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,12288,128,0.3054293261633979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,12288,32,0.020350222786267597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,64,0.05520888831880358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,12288,32,0.05617244376076592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,65536,1.1772168477376301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,16384,0.9574951595730252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,65536,4.020466698540582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,16384,1.2900222142537434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,16384,0.24848355187310112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,16384,1.3847689098782008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,12288,0.9485182232326932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,65536,5.995861477322048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,65536,5.252927992078993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,12288,0.183533337381151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,12288,0.9483280181884766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,10240,0.4475688934326172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,12288,1.0043244891696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,10240,0.7803760104709201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,10240,0.16063466336992052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,10240,0.8476355340745715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,8192,0.37275200419955784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,8192,0.12995999389224583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,8192,0.6580888960096571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,7168,0.5732133123609755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,8192,0.6879128879970975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,7168,0.2006115516026815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,6144,0.26446133189731175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,7168,0.54003021452162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,7168,0.6068373256259495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,6144,0.10412889056735569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,5120,0.22098933325873482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,6144,0.4616595374213324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,6144,0.5742684470282661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,5120,0.38720978630913633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,5120,0.09259021944469875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,4096,0.17730932765536836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,5120,0.4470977783203125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,4096,0.30759644508361816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,4096,0.37148533927069766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,3584,0.15883555677202013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,3584,0.2647075653076172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,3584,0.06484888659583198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,3584,0.3357715606689453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,3072,0.12921778361002603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,3072,0.22970310846964517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,3072,0.32845865355597603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,2560,0.10936977465947469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,4096,0.08076710833443536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,2560,0.19605422019958496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,2048,0.09053066704008315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,2560,0.33488088183932835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,2560,0.053454223606321544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,2048,0.1613617738087972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,2048,0.3281875451405843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,1536,0.0691831111907959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,3072,0.05890311135186089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,1536,0.12603910764058432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,1024,0.049843556351131864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,1536,0.34563732147216797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,1536,0.03906044363975525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,1024,0.093941330909729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,1024,0.032961779170566134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,768,0.03975288735495673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,1024,0.3354239993625217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,768,0.07541866434945001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,512,0.030068443881140813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,768,0.32830577426486546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,2048,0.046451555358039014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,512,0.061942219734191895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,512,0.028233776489893597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,512,0.37455021010504824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,256,0.023306666149033442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,256,0.05186400148603651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,256,0.38129509819878477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,256,0.025617778301239014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,128,0.020106666617923312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,768,0.030900445249345567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,128,0.04817422231038412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,10240,128,0.024937777055634394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,64,0.017954667409261067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,10240,32,0.018966222802797954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,64,0.048208889034059316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,10240,32,0.048488001028696694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,10240,128,0.3620159890916612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,65536,1.3598346710205078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,65536,2.9630275302463107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,16384,1.1835732989841037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,16384,1.1172000037299263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,16384,0.23010489675733778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,12288,0.4673413170708551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,65536,5.013937632242839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,16384,1.2244008382161458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,12288,0.16490666071573892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,12288,0.851725313398573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,65536,4.663875579833984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,12288,0.9148364596896701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,10240,0.37794134351942277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,10240,0.1370017793443468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,8192,0.3424933221605089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,10240,0.673708438873291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,8192,0.11356355084313287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,10240,0.7832329008314344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,8192,0.5708906915452745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,7168,0.25778577062818736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,8192,0.6335644192165798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,7168,0.10158577892515396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,7168,0.47230667538113064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,6144,0.21702310774061415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,7168,0.5575973192850748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,5120,0.17517956097920737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,6144,0.5478960143195258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,6144,0.5055475764804417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,5120,0.33851199679904515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,4096,0.13406311141120061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,5120,0.07947999901241727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,5120,0.4183759954240587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,4096,0.06429511308670044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,4096,0.26795466740926105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,3584,0.1293155617184109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,4096,0.4059448772006565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,6144,0.12376533614264594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,3072,0.11402044031355117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,3584,0.2335022290547689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,3584,0.05763733386993408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,3584,0.32487466600206166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,2560,0.09557777643203735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,3072,0.20270221763186982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,3072,0.051111999485227794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,3072,0.3176186614566379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,2048,0.07337155607011583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,2560,0.17078310913509795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,2560,0.04538311229811775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,2560,0.3215644359588623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,1536,0.05899555815590752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,2048,0.13973155286577013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,2048,0.32250934176974827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,1536,0.11016533109876846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,1536,0.0328764451874627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,1024,0.04088000125355191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,1536,0.3277448813120524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,1024,0.02738044493728214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,768,0.032708442873424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,1024,0.33126576741536456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,768,0.06595289044910006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,768,0.31518133481343585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,512,0.02403644389576382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,768,0.0236462222205268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,2048,0.03849244448873732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,512,0.053957333167394005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,256,0.01844888925552368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,512,0.022135999467637803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,512,0.32711466153462726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,256,0.044469333357281156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,128,0.017443555924627516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,1024,0.08292000161276923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,256,0.4002506732940674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,8192,128,0.2719564437866211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,64,0.018931556079122756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,128,0.020613332589467365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,8192,32,0.01959822244114346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,64,0.03779200050565932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,32,0.03735111157099406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,8192,256,0.020842666427294414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,8192,128,0.038557334078682795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,65536,1.1760187149047852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,65536,2.561192830403646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,16384,0.9533768759833442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,16384,1.03730042775472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,16384,1.1577449374728732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,65536,4.530510372585721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,16384,0.1989511118994819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,12288,0.3745608859592014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,65536,4.373040093315972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,12288,0.7598577605353461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,10240,0.3218844466739231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,12288,0.8712231318155924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,10240,0.12471732828352187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,10240,0.6225902239481608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,8192,0.2533537811703152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,12288,0.14213689168294272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,10240,0.7327057520548502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,8192,0.09845688607957627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,7168,0.21618221865759957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,8192,0.6577235327826606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,7168,0.4378808869255914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,7168,0.09024800194634332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,7168,0.5519030888875326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,8192,0.5194897651672363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,6144,0.37889954778883195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,6144,0.4606017536587185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,5120,0.14891821808285183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,6144,0.08207111226187812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,5120,0.31584265496995717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,5120,0.07256088654200236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,4096,0.11914932727813721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,6144,0.17858755588531494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,5120,0.39252887831793887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,3584,0.10860800080829197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,4096,0.05599288807974922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,4096,0.24846487575107148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,4096,0.32663554615444607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,3072,0.09447377920150757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,3584,0.05271288752555847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,3584,0.2147279977798462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,3584,0.324403550889757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,2560,0.07833066913816664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,3072,0.19268266359965006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,3072,0.046558221181233726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,3072,0.3701084454854329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,2560,0.15813777181837294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,2048,0.06564799944559734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,2560,0.04130311144722833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,2560,0.32643289036220974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,1536,0.05189955565664503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,2048,0.12979911433325872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,2048,0.035212443934546575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,2048,0.35661954349941677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,1536,0.10318844185935126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,1536,0.029010666741265193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,1536,0.327601777182685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,1024,0.07663910918765597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,768,0.027112889620992873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,1024,0.02442577812406752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,768,0.061816891034444175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,1024,0.42847288979424375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,512,0.02211644417709774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,768,0.3288559913635254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,768,0.021430222524536982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,512,0.019296000401178997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,256,0.016671111186345417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,512,0.32595822546217174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,1024,0.036066667901145086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,256,0.03874933388498094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,128,0.014583999911944071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,256,0.3203839990827772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,128,0.03460355599721273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,128,0.017605332864655387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,64,0.012450666891203986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,64,0.03399111164940728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,7168,32,0.012798222402731577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,512,0.049703111251195274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,7168,32,0.03354933195643955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,7168,256,0.019560888409614563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,7168,128,0.308680878745185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,65536,0.796711974673801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,65536,2.1640781826443143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,16384,0.5694453451368544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,16384,0.9514693154229058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,16384,0.1746631066004435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,12288,0.3446897665659587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,16384,1.103615125020345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,65536,4.385888841417101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,12288,0.6975884437561035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,12288,0.8334337870279948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,12288,0.13240267170800105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,10240,0.27405420939127606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,65536,4.250142415364583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,8192,0.2281902101304796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,10240,0.5779111120435927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,10240,0.7063760227627225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,8192,0.4784391191270616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,8192,0.0910595522986518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,7168,0.18815378348032633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,8192,0.5751546753777398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,7168,0.0803040001127455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,10240,0.1130595604578654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,6144,0.1592577828301324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,7168,0.515516440073649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,6144,0.3497626781463623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,6144,0.45593955781724715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,5120,0.138117339875963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,6144,0.0739368862575955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,7168,0.4057608975304498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,5120,0.2902586725023058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,5120,0.0656800005171034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,5120,0.5221920013427734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,4096,0.23041955629984537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,4096,0.052503112289640635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,4096,0.5125741958618164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,3584,0.09133155478371514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,4096,0.1009368896484375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,3584,0.19491733445061576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,3584,0.3948169019487169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,3072,0.08067999945746528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,3584,0.046181334389580615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,2560,0.07217955589294434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,3072,0.04126222266091241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,3072,0.33614667256673175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,2560,0.1461662186516656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,2048,0.052777777115503945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,2560,0.44844532012939453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,2048,0.11960089206695557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,2048,0.030140442980660334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,1536,0.045731554428736367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,2048,0.42078044679429794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,3072,0.169267561700609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,1536,0.09477244483100043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,1024,0.03136444422933791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,2560,0.03645777702331543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,1536,0.025928000609079998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,1536,0.39644267823961044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,768,0.026103999879625108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,1024,0.07050933440526326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,1024,0.021632888250880774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,1024,0.30332088470458984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,512,0.020095111595259774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,768,0.056536886427137584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,768,0.019472888774342008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,768,0.33365689383612734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,256,0.015248000621795654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,512,0.04404888881577385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,512,0.017602667212486267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,256,0.03517155514823066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,128,0.013421333498424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,512,0.4069111082288954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,256,0.016545777519543965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,128,0.030926220946841772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,6144,128,0.016368000043763053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,128,0.3876951005723741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,64,0.030622220701641504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,32,0.012096000214417776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,6144,32,0.030525333351559106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,6144,256,0.3103768825531006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,6144,64,0.011459555890825061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,65536,1.9704844156901042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,65536,0.7165830930074056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,16384,0.5245999760097927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,16384,0.8666986889309354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,16384,0.14994222588009304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,16384,1.0362346437242296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,12288,0.2836719883812798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,65536,3.716277228461372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,12288,0.1168408923678928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,12288,0.6371004316541884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,12288,0.7885120179918078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,10240,0.2644497818417019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,10240,0.10071021980709499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,10240,0.5292977756924099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,65536,3.957582261827257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,8192,0.20321867201063368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,10240,0.6610826916164821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,8192,0.08182844188478258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,7168,0.1817244423760308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,8192,0.43930400742424863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,8192,0.5431342124938965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,6144,0.15479289160834417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,7168,0.37144888771904844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,7168,0.07241600089603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,7168,0.4805697864956326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,5120,0.12230755223168267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,6144,0.06535910897784762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,6144,0.32400888866848415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,6144,0.4167288939158122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,4096,0.10055022107230292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,5120,0.058098667197757296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,5120,0.2674977779388428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,5120,0.4093804359436035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,4096,0.04484711090723673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,3584,0.09046222103966607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,4096,0.21133244037628174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,4096,0.33881423208448624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,3072,0.07783733473883735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,3584,0.04159644577238295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,3584,0.18257778220706514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,3584,0.4082195493910048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,3072,0.15650667084587946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,3072,0.03594311078389486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,2560,0.06456355253855388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,3072,0.3309582339392768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,2560,0.13245156076219347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,2560,0.03249511122703552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,2048,0.05057155423694187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,2560,0.3622844484117296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,2048,0.026439110438028973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,2048,0.11095466878679063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,1536,0.03965244359440274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,2048,0.38755732112460667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,1536,0.08752533462312485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,1536,0.022907555103302002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,1024,0.02514755560292138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,1536,0.37061511145697695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,1024,0.06436622142791748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,1024,0.01920977731545766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,768,0.02244088881545597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,1024,0.3871155579884847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,768,0.050380445188946195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,768,0.01776088939772712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,512,0.019824888971116807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,768,0.3697768847147624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,512,0.03991555505328708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,256,0.016192000773217943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,512,0.38031556871202254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,256,0.031552000178231135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,256,0.014873777826627096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,256,0.37410132090250653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,128,0.011166222393512726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,128,0.02798933453030056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,64,0.009735999835862054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,5120,128,0.32459823290507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,128,0.01516266663869222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,5120,32,0.01146311147345437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,64,0.027279999521043565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,5120,512,0.01593955523437924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,5120,32,0.027659555276234944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,65536,0.5170373386806911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,65536,1.3370320002237956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,16384,0.3461475637223985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,16384,0.7794008784823947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,16384,0.12781511412726507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,16384,0.9641350640190972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,12288,0.2448240121205648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,12288,0.5794897609286839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,65536,3.37589963277181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,10240,0.198799106809828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,12288,0.7539226743910047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,12288,0.09678666459189521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,65536,3.683075375027127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,10240,0.08368621932135688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,8192,0.16432355509863958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,10240,0.5034497578938802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,10240,0.6189991103278266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,7168,0.1433688931994968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,8192,0.06629688872231378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,8192,0.3974897861480713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,8192,0.507988452911377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,7168,0.06158488988876343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,6144,0.11801421642303467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,7168,0.33876532978481716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,7168,0.4470675786336263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,5120,0.09404622183905707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,6144,0.05646755297978719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,6144,0.2928248776329888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,6144,0.39191378487481016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,4096,0.08059733443790011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,5120,0.24344444274902344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,5120,0.04964800013436211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,4096,0.19346311357286242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,4096,0.03930933276812235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,4096,0.3264595667521159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,3584,0.0670142239994473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,3584,0.034154666794670954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,3072,0.0587715572781033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,3584,0.1659111049440172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,3584,0.32691823111640084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,3072,0.030343112018373277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,3072,0.14221155643463135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,2560,0.0495395561059316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,3072,0.38330311245388454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,5120,0.332590209113227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,2560,0.12018667327033149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,2048,0.044216887818442456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,2560,0.2854959964752197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,2560,0.02738933265209198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,1536,0.03292444348335266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,2048,0.09859200318654378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,2048,0.023315555519527856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,2048,0.3218008942074246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,1024,0.023735110958417256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,1536,0.01873333255449931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,1536,0.0790826678276062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,1536,0.32279909981621635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,768,0.01946311030122969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,1024,0.0162773331006368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,1024,0.3209555678897434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,768,0.04612266686227587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,512,0.013955555028385587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,768,0.016011555989583332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,512,0.03534844517707825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,768,0.41112534205118817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,512,0.34818222787645126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,256,0.011197333534558615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,512,0.01423288881778717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,256,0.027787556250890095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,1024,0.057160000006357826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,128,0.009672000176376766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,256,0.013120888835854001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,256,0.4554462432861328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,128,0.025065778030289546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,64,0.009322666459613377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,4096,128,0.013431111143694984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,4096,128,0.33737778663635254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,4096,32,0.009767110976907942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,64,0.02457511093881395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,4096,32,0.024088889360427856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,65536,1.3481644524468317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,65536,0.4595022201538086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,16384,0.2733173370361328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,16384,0.7584115664164225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,16384,0.9314746856689453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,65536,3.0589343176947703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,16384,0.11533599429660374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,12288,0.21553511089748809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,12288,0.0885288914044698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,12288,0.5417440202501085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,10240,0.15776711040072972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,65536,3.5166142781575522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,12288,0.7027084562513563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,8192,0.12843110826280382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,10240,0.07467999723222521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,10240,0.4555066426595052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,10240,0.5959822336832682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,8192,0.06186222367816501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,7168,0.11170844237009685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,8192,0.4078533384535048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,8192,0.48709689246283633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,6144,0.09411022398206924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,7168,0.3221457799275716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,7168,0.05592177973853218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,7168,0.4331066608428955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,5120,0.08056977722379896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,6144,0.05041155550214979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,6144,0.27105334069993764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,6144,0.3715173403422038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,4096,0.06551200151443481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,5120,0.04254755708906385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,5120,0.22708267635769316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,5120,0.4125795629289415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,4096,0.03498933381504483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,4096,0.18331200546688506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,3584,0.05692355500327217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,4096,0.3228222264183892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,3072,0.048193779256608754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,3584,0.03132177723778619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,3584,0.3363813294304742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,3072,0.13557867209116617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,2560,0.041356444358825684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,3072,0.02810844447877672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,3072,0.3708817693922255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,2560,0.11391554938422309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,2048,0.0324942237801022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,2560,0.3288408915201823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,2560,0.025099555651346844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,2048,0.09273333019680446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,2048,0.02103999919361538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,1536,0.02665688925319248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,2048,0.3213262293073866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,3584,0.15470488866170248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,1536,0.018785778019163344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,1536,0.0744488901562161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,1536,0.3236444526248508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,1024,0.05400000015894572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,1024,0.015665777855449252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,768,0.015487111277050443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,1024,0.39614934391445583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,768,0.0439208878411187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,512,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,768,0.33240000406901044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,768,0.014508444401952954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,512,0.033907555871539645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,256,0.009516444471147325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,512,0.013506666653686099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,512,0.3178657690684001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,256,0.025733333494928148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,128,0.008024888734022776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,256,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,128,0.02304711110062069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,256,0.4146240022447374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,64,0.007598222129874759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3584,128,0.30015465948316783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3584,128,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,32,0.007658667034573025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,64,0.022727110319667395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3584,32,0.022290666898091633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3584,1024,0.01963644391960568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,65536,0.85971376630995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,16384,0.28238312403361004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,65536,0.4467662175496419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,16384,0.7563244501749674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,16384,0.9262018203735352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,16384,0.11043466462029351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,12288,0.1718613306681315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,65536,2.878523508707682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,12288,0.5172133445739746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,12288,0.7029822137620715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,65536,3.5074524349636502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,10240,0.15107022391425237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,10240,0.43105244636535645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,10240,0.5944773356119791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,10240,0.07237511210971408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,8192,0.12182666195763482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,12288,0.08511288960774739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,7168,0.10936178101433648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,8192,0.05901244613859388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,8192,0.35330666436089414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,8192,0.49657689200507266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,7168,0.05369155605634054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,6144,0.0919884443283081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,7168,0.30657511287265354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,7168,0.43080888854132754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,5120,0.07006311416625977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,6144,0.26433155271742076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,6144,0.04814755585458544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,6144,0.4242400063408746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,5120,0.041309333509869046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,4096,0.05900444587071737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,5120,0.3283155494266086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,5120,0.21435556146833631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,3584,0.05025510986646017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,4096,0.17425510618421766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,4096,0.03307822346687317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,4096,0.3895271089341905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,3072,0.04237066706021627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,3584,0.02930844492382473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,3584,0.14654844337039524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,3584,0.3384302192264133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,2560,0.03692444496684604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,3072,0.026362667481104534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,3072,0.12903466489579943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,3072,0.33287644386291504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,2048,0.03133599956830343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,2560,0.10910222265455459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,2560,0.023357333408461675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,2048,0.08833422263463338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,2560,0.40719021691216367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,1536,0.0248391115003162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,2048,0.324689785639445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,1536,0.06864088773727417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,1536,0.016978666186332703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,1024,0.016338666280110676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,1536,0.33143289883931476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,1024,0.05111644334263272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,1024,0.014479999740918478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,768,0.014674666855070325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,1024,0.40140533447265625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,768,0.32829509841071236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,512,0.012364444633324942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,768,0.013853333062595792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,2048,0.020000888241661918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,512,0.0325893329249488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,256,0.010213333699438307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,512,0.013083555632167392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,256,0.024267555938826665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,512,0.414722654554579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,128,0.010025777750545079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,256,0.3482951058281793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,256,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,128,0.02199199961291419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,3072,128,0.012376888758606382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,64,0.009510222408506606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,3072,32,0.009741333623727163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,3072,128,0.3269591066572401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,64,0.02136444383197361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,32,0.02129688858985901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,65536,0.7807377709282769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,3072,768,0.04010399844911363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,16384,0.20147289170159233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,65536,0.48749690585666233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,16384,0.648563543955485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,16384,0.099181334177653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,16384,0.9161333507961698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,12288,0.15312177605099148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,65536,2.6496319240993924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,12288,0.0767093300819397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,12288,0.4876560105217828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,10240,0.12283555666605632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,12288,0.6748204231262207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,10240,0.06538133488761054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,10240,0.410079108344184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,8192,0.09700444671842788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,65536,3.380377875434028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,10240,0.6219706535339355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,7168,0.08591022094090779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,8192,0.33213334613376194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,8192,0.052925331725014575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,8192,0.46478668848673504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,6144,0.07059199942482842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,7168,0.048539555735058255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,7168,0.29119910134209526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,7168,0.4135848946041531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,6144,0.24970756636725533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,6144,0.04402044415473938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,5120,0.05708977911207411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,6144,0.41698667738172746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,5120,0.20806666215260824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,5120,0.03744533326890733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,5120,0.38750399483574766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,4096,0.045145776536729604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,4096,0.03045155604680379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,3584,0.04190222091144986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,4096,0.38354934586419004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,3584,0.1420311133066813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,3584,0.36070044835408527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,3072,0.037765334049860634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,3072,0.12084710597991943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,4096,0.16511466768052843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,3072,0.36844534344143337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,3072,0.02392888896995121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,2560,0.03028977910677592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,3584,0.02760266595416599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,2560,0.10196533468034531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,2048,0.027615111735132005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,2560,0.34318844477335614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,2048,0.018889778190188937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,1536,0.021351110604074266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,2048,0.34633776876661515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,1536,0.06301155355241564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,1536,0.3305342197418213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,1536,0.015547555353906421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,1024,0.015639111399650574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,1024,0.047163556019465126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,1024,0.33510933982001406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,2560,0.02132266594303979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,768,0.012487110992272695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,2048,0.08289955721961127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,768,0.03785066803296407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,512,0.011164444188276926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,768,0.33451199531555176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,768,0.01312444441848331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,512,0.02917155623435974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,256,0.009483555952707926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,512,0.012433778080675336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,512,0.3197733296288384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,256,0.022986667023764715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,256,0.01181333346499337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,128,0.008188444707128737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,1024,0.013869333598348828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,256,0.41255821122063535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,128,0.02085955606566535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,64,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2560,32,0.008031110796663497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2560,128,0.34016889995998806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2560,128,0.01144266708029641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,64,0.01998488936159346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2560,32,0.019691555036438834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,65536,0.5814977751837836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,16384,0.1686506668726603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,65536,0.33846577008565265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,16384,0.6128355662027994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,16384,0.8508417871263293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,16384,0.08905955817964341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,12288,0.11887821886274551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,12288,0.46164798736572266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,65536,2.4562301635742188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,12288,0.6470026440090603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,10240,0.09995555877685547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,12288,0.0674720009167989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,10240,0.058752887778811984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,8192,0.07742755280600654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,10240,0.38698400391472715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,10240,0.5474470986260308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,65536,3.2293128967285156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,8192,0.04673333300484551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,7168,0.07292444176144071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,8192,0.4516400231255426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,7168,0.2739493317074246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,7168,0.4017911222245958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,6144,0.06068355507320828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,6144,0.23381243811713326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,6144,0.3517875671386719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,5120,0.05194311009513008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,6144,0.03863111138343811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,8192,0.31430578231811523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,5120,0.1966693268881904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,4096,0.04031022058592902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,5120,0.03250844279925028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,5120,0.3344213432735867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,7168,0.04251377781232198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,4096,0.02693066663212246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,3584,0.037124445041020714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,4096,0.33498043484157985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,3584,0.13365422354804143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,3584,0.024354666471481323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,3072,0.03222577770551046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,3584,0.4154942300584581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,3072,0.11490755610995823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,2560,0.0273635552989112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,3072,0.32984089851379395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,3072,0.02181688944498698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,2560,0.09630578094058567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,2560,0.3695102267795139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,4096,0.155114663971795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,2560,0.019122666782803006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,2048,0.02381688853104909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,2048,0.016952888833151925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,1536,0.018146667215559218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,2048,0.42493422826131183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,1536,0.05920266442828708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,1536,0.014854222536087036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,1536,0.4079271157582601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,1024,0.013906665974193148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,1024,0.044086221191618175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,1024,0.3751102288564046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,768,0.011844444606039258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,768,0.03487199876043532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,768,0.3698737886216905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,512,0.009670221971140968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,768,0.012420444852775998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,2048,0.07868178023232354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,512,0.027531555957264368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,256,0.007692444655630324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,512,0.01144444445768992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,256,0.021298666795094807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,256,0.010806221928861407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,256,0.31548799408806694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,128,0.006615110983451207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,128,0.01888799998495314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,128,0.3085413244035509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,128,0.010752889017264048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,64,0.007279111279381647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,2048,32,0.007290666302045186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,2048,1024,0.012854222622182635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,64,0.018216888109842937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,2048,32,0.018280888597170513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,65536,0.4204622109731038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,2048,512,0.5548204316033257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,16384,0.11947555012173122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,65536,0.3224959903293186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,16384,0.5745795567830404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,16384,0.8161342408921984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,16384,0.0762266649140252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,12288,0.09067822164959377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,65536,2.2920843760172525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,12288,0.4342888991038005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,10240,0.08643288744820489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,12288,0.05722666449016995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,10240,0.3635404374864366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,10240,0.5250008900960287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,65536,3.0905804104275174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,8192,0.06748355759514703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,10240,0.051177776522106595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,8192,0.2947022120157878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,8192,0.04048622316784329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,12288,0.6224577691819932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,7168,0.2583351135253906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,7168,0.38314133220248753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,6144,0.05279911226696438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,7168,0.03700266612900628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,6144,0.22433333926730684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,6144,0.032820443312327065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,8192,0.4320231013827854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,5120,0.0425582230091095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,6144,0.44785865147908527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,5120,0.1838515599568685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,5120,0.029471112622155085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,4096,0.032391998502943255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,5120,0.34388266669379336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,4096,0.023112888137499493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,4096,0.14643466472625732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,3584,0.029683556821611192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,4096,0.33782045046488446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,3584,0.02091644373204973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,3584,0.12626666492886013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,3584,0.33580533663431805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,7168,0.05891288651360405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,3072,0.018949333164427016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,3072,0.10755466752582127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,2560,0.022459555003378127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,3072,0.3314710987938775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,2560,0.016930666234758165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,2560,0.09095822440253364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,2560,0.34172712432013613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,2048,0.07467555337482028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,2048,0.3293902344173855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,3072,0.02658044464058346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,1536,0.013992889059914483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,2048,0.015174221661355762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,1536,0.05746755335066053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,1024,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,1536,0.013266666895813413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,1536,0.3270924356248644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,1024,0.03970400161213345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,768,0.009558222360081142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,1024,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,2048,0.018041777941915724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,1024,0.4075271023644342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,768,0.03249511122703552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,768,0.011026666396194033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,512,0.008231999973456064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,768,0.3276657793256971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,512,0.025789333714379206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,256,0.006519111080302133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,512,0.010672888822025724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,512,0.363630215326945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,256,0.0196115556690428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,128,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,256,0.010044444766309526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,256,0.35337599118550617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,128,0.017249777913093567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,64,0.005929777605666055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1536,128,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1536,128,0.3525377909342448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1536,32,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,64,0.016622222132152982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1536,32,0.016876444220542908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,65536,0.3185262150234646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,65536,0.2931777901119656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,16384,0.12598578135172525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,16384,0.5594142278035482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,16384,0.0643813345167372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,12288,0.06589688857396443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,65536,2.1186372968885636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,16384,0.7727093166775174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,12288,0.40520622995164657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,10240,0.05179199907514784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,12288,0.04924177792337206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,12288,0.5951093567742242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,10240,0.33948267830742734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,10240,0.043471111191643604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,10240,0.5172826449076334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,8192,0.2750391165415446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,8192,0.03399822115898132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,7168,0.03853866789076064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,8192,0.4912106725904677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,7168,0.2412559986114502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,8192,0.04353866643375821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,6144,0.03253600001335144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,7168,0.4632844395107693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,6144,0.20805154906378853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,6144,0.44270579020182294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,5120,0.02860622273551093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,5120,0.17350044515397814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,65536,2.9557323455810547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,7168,0.030269334713617962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,4096,0.022817777262793645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,5120,0.02420444455411699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,5120,0.41721333397759336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,4096,0.13555288314819336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,3584,0.021197333931922913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,4096,0.33281156751844615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,6144,0.027331555883089702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,4096,0.019897777173254225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,3072,0.018759111563364666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,3584,0.018193778064515855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,3584,0.11769955688052708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,3584,0.3232755661010742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,2560,0.01587466730011834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,3072,0.016351110405392118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,3072,0.10013688935173883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,3072,0.3275502257876926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,2048,0.014137778017255994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,2560,0.08414844671885173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,2560,0.014856000741322836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,2560,0.3196622265709771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,1536,0.01090844472249349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,2048,0.0700942211680942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,2048,0.3220844533708361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,1536,0.051649779081344604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,1024,0.008643555144468943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,1536,0.011697777443461947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,1536,0.3297617700364855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,1024,0.036305778556399874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,1024,0.010696888797812991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,1024,0.3249724441104465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,768,0.03027555677625868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,768,0.010178666975763109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,768,0.33218489752875435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,512,0.0064737776087390045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,2048,0.013077333569526672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,512,0.02404888967672984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,256,0.005585777676767773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,512,0.33962310685051816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,512,0.009756444229020013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,256,0.01843644513024224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,128,0.004578666554556953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,768,0.00738844441043006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,128,0.016158221496476066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,256,0.3339546786414252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,64,0.004899555610285865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,128,0.00925688859489229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,1024,128,0.27411643664042157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,1024,32,0.0052177777720822235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,64,0.015497777197096082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,1024,32,0.015145777000321282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,65536,0.22887910736931694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,65536,0.2797635661231147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,16384,0.06714133421579997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,16384,0.5240791108873155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,1024,256,0.009315555294354757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,65536,2.056531482272678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,16384,0.777982234954834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,12288,0.05048977666431003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,16384,0.06358310911390516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,12288,0.3962639967600505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,12288,0.5945164362589518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,65536,2.9466063181559243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,12288,0.04877510997984144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,10240,0.044124444325764976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,10240,0.04401600029733446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,8192,0.03519377774662442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,10240,0.33194221390618217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,10240,0.5026693344116211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,8192,0.26772978570726186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,8192,0.03338399860594008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,7168,0.029986666308508977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,8192,0.40842488076951766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,7168,0.03033155534002516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,7168,0.2360746595594618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,6144,0.02588711182276408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,7168,0.37709689140319824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,6144,0.026901332868470088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,5120,0.022487110561794702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,6144,0.20446133613586426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,6144,0.33412888314988876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,4096,0.01808888879087236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,5120,0.024491555160946313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,5120,0.1684631109237671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,4096,0.13256800174713135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,3584,0.01593777702914344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,4096,0.3423173427581787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,3584,0.11457155810462104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,3584,0.33475910292731387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,3584,0.01793066660563151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,3072,0.014226666755146451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,3072,0.09776088926527236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,3072,0.33080710305107963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,2560,0.012688889271683164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,3072,0.016255999604860943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,5120,0.373839987648858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,4096,0.019905777441130746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,2048,0.011324444578753578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,2560,0.08096800247828166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,2560,0.014835556348164877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,2560,0.3256266646915012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,1536,0.00905955582857132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,2048,0.06803644365734525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,2048,0.013219555219014486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,2048,0.32721956570943195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,1024,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,1536,0.050916443268458046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,1536,0.011786667009194693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,1536,0.32328799035814076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,1024,0.036552889479531184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,1024,0.3212053245968289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,768,0.029492444462246362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,768,0.010053333308961656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,512,0.005972444597217772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,768,0.327056884765625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,512,0.023402666052182514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,512,0.3187297715081109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,512,0.009669333696365356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,256,0.004898666507667965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,256,0.01755466726091173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,1024,0.01069155583779017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,256,0.32579021983676487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,256,0.009047110875447592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,128,0.004531555705600315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,128,0.015113777584499784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,64,0.003859555555714501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,768,128,0.3244711028205024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,32,0.005917333480384614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,64,0.014429334137174817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,768,32,0.014490667316648694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,65536,0.17440266079372832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,768,768,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,65536,0.1898862255944146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,16384,0.04966488811704847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,768,128,0.009125333693292405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,16384,0.5041982332865397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,16384,0.7376302083333334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,16384,0.05226844549179077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,12288,0.03866666555404663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,65536,1.978859583536784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,12288,0.3815324571397569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,10240,0.03305955396758185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,12288,0.04046666622161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,12288,0.5663670963711208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,10240,0.03636622097757127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,8192,0.02428177826934391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,65536,2.811492496066623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,10240,0.48097954856024844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,8192,0.026778666509522334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,7168,0.02180266711446974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,8192,0.2590471108754476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,8192,0.3950897852579753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,7168,0.024082667297787134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,7168,0.22808533244662813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,6144,0.01871466636657715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,7168,0.4143928951687283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,10240,0.3198640081617567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,6144,0.19639289379119873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,5120,0.015782222151756287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,6144,0.35887734095255536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,6144,0.022014222211307947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,5120,0.16363288296593562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,5120,0.02032533288002014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,4096,0.013576889203654395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,5120,0.3642417854732937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,4096,0.016173332929611206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,4096,0.1279333300060696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,3584,0.011878222227096558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,4096,0.3559599982367621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,3584,0.11069599787394206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,3072,0.010945777926180096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,3584,0.3753697872161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,3072,0.0942408906088935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,3072,0.3472542232937283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,3072,0.01421511173248291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,2560,0.009450666606426239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,2560,0.07908355527453952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,2560,0.34456798765394425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,2048,0.008259555531872643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,2560,0.01277599980433782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,3584,0.015202666322390238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,1536,0.007438222567240397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,2048,0.011409777734014722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,2048,0.06405866808361478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,2048,0.32724444071451825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,1536,0.010234666367371878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,1024,0.006173333360089197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,1536,0.0485226677523719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,1536,0.32190489768981934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,768,0.005517333332035277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,1024,0.03365155392222934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,1024,0.32536888122558594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,1024,0.00979200005531311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,768,0.02816177739037408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,512,0.004691555682155821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,768,0.009353777600659264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,768,0.31906400786505806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,256,0.004223111189073986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,512,0.022720000810093347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,512,0.00869066682126787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,256,0.016746666696336534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,128,0.003921777837806278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,256,0.008360889222886827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,256,0.3189217779371473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,512,0.49733956654866535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,64,0.003975999852021535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,128,0.014456000592973499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,512,32,0.003547555456558863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,512,128,0.3032337824503581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,32,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,65536,0.10902221997578938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,512,128,0.008344000412358178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,512,64,0.013839999834696451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,16384,0.037668443388409086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,65536,0.18578044573465982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,65536,1.8992862701416016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,16384,0.4854719903733995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,12288,0.028023110495673284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,16384,0.7379662195841471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,16384,0.05087466703520881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,12288,0.039592888620164655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,10240,0.021783999270863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,12288,0.36756266487969297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,12288,0.5637244648403591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,65536,2.793368869357639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,8192,0.01734755602147844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,10240,0.3081128862169054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,10240,0.035997334453794695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,10240,0.47735553317599827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,8192,0.02688088847531213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,8192,0.24808798895941842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,8192,0.3931146727667914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,7168,0.2190097835328844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,7168,0.023966223001480103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,6144,0.013597333596812354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,7168,0.3507573339674208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,6144,0.022164444128672283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,5120,0.012048888537618848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,6144,0.18872800138261583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,6144,0.34291733635796445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,5120,0.020247111717859905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,5120,0.15645599365234375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,4096,0.010494222243626913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,5120,0.3329546716478136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,7168,0.015834665960735746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,3584,0.009447111023796929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,4096,0.016174222032229107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,4096,0.12295733557807074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,4096,0.32484534051683217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,3072,0.008718222379684448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,3584,0.015186667442321777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,3584,0.10627555847167969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,3584,0.3171146710713704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,2560,0.007686222592989604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,3072,0.08990666601392959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,3072,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,3072,0.33104801177978516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,2560,0.07546755340364245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,2048,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,2560,0.012814222110642327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,2560,0.3672835561964247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,2048,0.061659554640452065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,1536,0.005929777605666055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,2048,0.3853600025177002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,1536,0.04557866520351834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,1536,0.01036800030204985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,1024,0.005223111146026188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,1536,0.4020995563930935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,1024,0.032454222440719604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,1024,0.3548444377051459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,1024,0.009338666995366415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,2048,0.01148799972401725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,768,0.027165333429972332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,768,0.009007111191749573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,768,0.3831813335418701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,512,0.004232888834344016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,512,0.02165422174665663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,512,0.3497226768069797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,512,0.00866844422287411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,256,0.004006222304370669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,768,0.004576000074545543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,256,0.01581333412064446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,128,0.0035031110876136353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,256,0.33365599314371747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,128,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,64,0.0031795555518733132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,128,0.00835111074977451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,256,128,0.30557865566677517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,64,0.01311555587583118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,65536,0.08775555425220066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,256,256,0.00847822262181176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,256,32,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,256,32,0.013018666870064206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,16384,0.02888622217708164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,65536,0.1837599939770169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,65536,1.8572711944580078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,16384,0.47469245062934023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,12288,0.02086755633354187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,16384,0.7340648969014486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,16384,0.05230222145716349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,12288,0.041272001134024724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,12288,0.3595511118570964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,10240,0.015356444650226168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,12288,0.5623146692911783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,10240,0.3011973434024387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,10240,0.03752533263630337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,8192,0.013190222283204397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,10240,0.4767573144700792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,8192,0.027707555227809485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,7168,0.011727111207114326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,8192,0.24370845158894858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,8192,0.3926604323916965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,65536,2.789475546942817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,6144,0.010460444622569613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,7168,0.02511022157139248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,7168,0.2139813370174832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,7168,0.3499013317955865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,5120,0.009135999613338048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,6144,0.022836445106400385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,6144,0.18513777520921496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,6144,0.33570843272738987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,4096,0.00868977771864997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,5120,0.15453599558936224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,5120,0.021003555920388963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,5120,0.3205431037478977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,3584,0.007601777712504069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,4096,0.12093867195977105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,4096,0.016914667354689706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,4096,0.32940443356831867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,3072,0.0069208890199661255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,3584,0.10367021958033244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,3584,0.015561777684423657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,3584,0.2965128951602512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,2560,0.006623111251327727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,3072,0.08780266841252644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,3072,0.014458666245142618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,3072,0.315591123369005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,2560,0.07370577918158637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,2048,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,2560,0.32092176543341744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,2048,0.059839109579722084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,1536,0.00489155575633049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,2048,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,2048,0.3134053283267551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,1536,0.010392000277837118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,1024,0.004530666602982415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,1536,0.32900532086690265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,1024,0.03161599900987413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,1024,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,1024,0.3354186746809218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,768,0.0038942222793896994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,2560,0.01316177762216992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,768,0.026498667067951624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,512,0.003607999947335985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,768,0.32068710856967503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,768,0.0094035557574696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,512,0.021010667085647583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,1536,0.04478044311205546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,256,0.0032755554550223877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,512,0.008708444734414419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,512,0.3394346767001682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,256,0.015440000428093804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,128,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,256,0.00833155545923445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,256,0.379530668258667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,128,0.013268444273206921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,64,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4096,128,128,0.008357333640257517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,128,32,0.0032604444358083936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,64,0.012778666284349231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,128,32,0.012458666331238218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,65536,0.08532088994979858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,16384,0.027120888233184814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,16384,0.47471290164523655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,12288,0.01660977800687154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,12288,0.3594382339053684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,10240,0.014567999376191033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,8192,0.01221244451072481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,65536,1.8571671379937067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,10240,0.3014604515499539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,7168,0.011160000330872007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,6144,0.010008888940016428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,6144,0.18485689163208008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,5120,0.008681777450773451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,7168,0.21426133314768472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,4096,0.007848000360859765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,5120,0.15405244297451443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,3584,0.0069422221018208405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,4096,0.1206479999754164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,8192,0.24348889456854927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,3072,0.006726222319735422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,3584,0.10371466477711995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,2560,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,2048,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,3072,0.08793333503935073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,1536,0.004604444321658877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,2048,0.059960001044803195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,1024,0.004087111188305749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,1536,0.04436711139149136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,768,0.0038773334688610504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,1024,0.031581332286198936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,768,0.02641511129008399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,512,0.003546666767862108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,256,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,512,0.02088711162408193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,128,0.0031964443624019623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,256,0.015220445063379077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,64,0.002924444360865487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,128,0.013096888860066732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,64,32,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,2560,0.0735751125547621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,64,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,64,32,0.012422222230169507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,16384,0.025207999679777358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,16384,0.474272886912028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,12288,0.013285333083735572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,12288,0.3592844539218479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,10240,0.011414222419261932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,10240,0.300963560740153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,65536,1.856387456258138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,8192,0.010020444790522257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,7168,0.009381333159075843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,8192,0.24353954527113172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,7168,0.2141253285937839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,6144,0.008709333009190029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,5120,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,6144,0.1847511132558187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,4096,0.010013333625263637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,5120,0.15351288848453098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,3584,0.009117333425415887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,4096,0.12020444869995117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,3072,0.0083155557513237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,3584,0.10396177901162042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,2560,0.0074240002367231585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,3072,0.08859911229875352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,2048,0.005221333354711533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,2560,0.07387200329038832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,2048,0.05977866384718153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,1536,0.04415555463896858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,1024,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,1024,0.03169688913557265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,768,0.0038648889296584656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,768,0.026457778281635706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,512,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,65536,0.0861404471927219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,256,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,512,0.020969778299331665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,128,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,256,0.015189333094490899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,64,0.0029146667155954572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,128,0.012823999755912356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,32,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,64,0.012084444363911947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4096,32,32,0.01238844460911221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4096,32,1536,0.005014222115278244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,16384,2.9379013909233938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,16384,3.0185936821831594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,16384,0.7354506916469998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,16384,3.041631062825521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,12288,2.1527972751193576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,12288,2.1481865776909723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,12288,0.5475466516282823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,10240,2.1587369706895614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,12288,2.5202969445122614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,10240,1.7334221733940973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,10240,0.4484071201748318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,8192,1.483909289042155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,8192,1.4859679539998372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,10240,1.8968772888183594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,8192,0.3663644525739882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,8192,1.3569066789415147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,7168,1.3484267128838434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,7168,1.1650809182061088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,7168,0.32351554764641655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,6144,1.1476817660861545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,7168,1.0971031188964844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,6144,1.0025741789076064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,6144,0.2872462272644043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,5120,0.6461048656039768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,6144,1.1654764811197917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,5120,0.8234453201293945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,5120,0.2758951187133789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,5120,0.7877626419067383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,4096,0.5371573236253526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,4096,0.20512178209092882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,4096,0.6684213214450412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4096,128,128,0.302689790725708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,3584,0.4737253189086914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,4096,0.6552542050679525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,3584,0.5811786651611328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,3584,0.22805333137512207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,3584,0.580800904168023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,3072,0.5000391006469727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,3072,0.15889155864715576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,2560,0.3226897716522217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,3072,0.5277457767062718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,3072,0.4094524383544922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,2560,0.14978755844963923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,2560,0.42457066641913516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,2048,0.2649795479244656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,2560,0.45734135309855145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,2048,0.3459164566463894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,1536,0.2281866603427463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,2048,0.14805955357021755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,1536,0.2763502332899305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,1536,0.1011928915977478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,1536,0.3277964327070448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,1024,0.1427475611368815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,1024,0.08298489120271471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,1024,0.2085368898179796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,1024,0.3391244411468506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,768,0.12445599502987331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,2048,0.38057245148552793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,768,0.07459200090832181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,768,0.1681413385603163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,768,0.3240728908114963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,512,0.15038843949635824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,512,0.06694222158855863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,256,0.06330666939417522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,512,0.3981502321031358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,256,0.1252773337894016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,128,0.053631110323799976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,256,0.0603804455863105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,128,0.12021688620249431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,512,0.0860862202114529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,128,0.3802737659878201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,65536,128,0.05669777923160129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,64,0.048453334305021495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,65536,32,0.04874133401446872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,64,0.12129510773552789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,65536,32,0.12184354994032119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,65536,256,0.32768355475531685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,65536,0.821815119849311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,65536,2.974975162082248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,16384,0.9145920011732313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,16384,0.8956089019775391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,65536,3.961249033610026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,16384,0.8808302349514432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,16384,0.19459199905395508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,12288,0.452822208404541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,12288,0.6397600173950195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,12288,0.16294666131337485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,12288,0.6810995737711588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,10240,0.3736862341562907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,10240,0.12858311335245767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,10240,0.5655991236368815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,8192,0.28773776690165204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,10240,0.5741919941372341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,8192,0.442213323381212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,8192,0.1049964427947998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,65536,3.6456887986924915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,7168,0.22483200497097441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,8192,0.5127813551161025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,7168,0.3671368757883708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,7168,0.09323288997014363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,7168,0.4138160016801622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,6144,0.3175528844197591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,6144,0.387523545159234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,6144,0.08374044630262588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,5120,0.16649422380659315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,5120,0.2649848990970188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,6144,0.22268001238505045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,5120,0.4392293294270833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,5120,0.07271377907858954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,4096,0.1440826654434204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,4096,0.06087377998563978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,4096,0.21384799480438232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,3584,0.13159467114342585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,4096,0.42834311061435276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,3584,0.05448888738950094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,3584,0.18581955962710908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,3072,0.11851733260684544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,3584,0.38451822598775226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,3072,0.0497555567158593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,3072,0.1625866625044081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,2560,0.09450578027301365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,3072,0.3788915475209554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,2560,0.04395822352833218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,2560,0.13985333177778456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,2048,0.07701333363850911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,2560,0.3683626651763916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,2048,0.11604444185892741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,2048,0.038542220989863075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,1536,0.055989331669277616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,2048,0.35460623105367023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,1024,0.04091466797722711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,1536,0.09291644228829278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,1536,0.030822220775816176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,1024,0.02463111115826501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,1024,0.0703528920809428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,1536,0.41742221514383954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,768,0.03222222129503886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,1024,0.3263271119859484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,512,0.024167999625205994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,768,0.02238222294383579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,768,0.05743822124269274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,512,0.04745599958631727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,768,0.3912115626864963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,256,0.018426666657129925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,512,0.020119110743204754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,512,0.3391244411468506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,128,0.017383999294704862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,256,0.04107466671201918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,256,0.01869244376818339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,256,0.3233395682440864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,64,0.01918577816751268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,128,0.03801688883039687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,16384,128,0.018225777480337355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,16384,32,0.01961599952644772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,64,0.037476443582110934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,16384,128,0.3731999927096897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,16384,32,0.03638755612903171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,65536,0.684383127424452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,16384,0.6335688696967231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,65536,2.201311959160699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,16384,0.7184933556450738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,16384,0.7708569102817111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,16384,0.1471831136279636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,65536,3.2779511345757375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,12288,0.33882755703396267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,12288,0.1163368887371487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,12288,0.5952000088161892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,65536,2.938167995876736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,10240,0.4311066733466254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,10240,0.5063546498616537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,10240,0.10062044196658665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,12288,0.5233955383300781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,8192,0.23318844371371797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,10240,0.27788710594177246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,8192,0.35487998856438535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,8192,0.08112533224953546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,7168,0.18750400013393828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,8192,0.408494234085083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,6144,0.154548446337382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,7168,0.07522577709621854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,7168,0.2960275544060601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,7168,0.4049839973449707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,6144,0.06670133272806804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,5120,0.1405173275205824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,6144,0.3355040020412869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,5120,0.21581155723995635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,5120,0.05831199884414673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,4096,0.11763111750284831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,5120,0.40818132294548887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,4096,0.17609155178070068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,4096,0.04923644330766466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,4096,0.33979378806220156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,3584,0.1017306645711263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,3584,0.04302755660480923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,3584,0.15444000562032065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,3072,0.08613866567611694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,3584,0.3262897862328423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,6144,0.25630489985148114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,2560,0.07230489121543036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,3072,0.03908088803291321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,3072,0.1345920032925076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,3072,0.32499554422166615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,2560,0.11526400513119167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,2048,0.05457955598831177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,2560,0.03451555636194017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,2560,0.3628728919559055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,2048,0.09603999720679389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,2048,0.029025776518715754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,1536,0.04229066769282023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,2048,0.3692311180962457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,1536,0.07730222410625882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,1536,0.02446844511561924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,1024,0.029927998781204224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,1536,0.37035287751091855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,1024,0.0578133331404792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,768,0.02474133339193132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,1024,0.019768888751665752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,1024,0.37675200568305117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,768,0.046579556332694165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,512,0.019458666443824768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,768,0.01832444469134013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,768,0.3461919890509711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,512,0.03803911142879062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,256,0.014955555399258932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,512,0.016591999265882704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,512,0.35222045580546063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,256,0.032347556617524885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,128,0.01351999988158544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,256,0.3673875596788194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,128,0.03051111102104187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,128,0.015207111835479736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,64,0.011524444652928246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,12288,128,0.3364826573265924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,12288,32,0.012059555285506778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,64,0.030629333522584703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,12288,32,0.03017510970433553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,12288,256,0.015697777271270752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,65536,0.5401466687520345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,65536,1.7335617277357314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,16384,0.39355821079678005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,16384,0.6339359813266331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,16384,0.12950666745503744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,65536,2.7280826568603516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,16384,0.7026151021321615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,12288,0.260879119237264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,12288,0.12184000015258789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,12288,0.4665866427951389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,10240,0.2582204341888428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,12288,0.5437306827969021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,65536,2.6051262749565973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,10240,0.3866639931996663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,10240,0.08910844723383586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,8192,0.20462932851579455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,10240,0.45385154088338214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,8192,0.3152204354604085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,8192,0.3786764409806993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,7168,0.16492088635762533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,7168,0.06390044424268934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,7168,0.26503554979960126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,7168,0.33884888225131565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,6144,0.22746488783094618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,6144,0.32769155502319336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,6144,0.05790577994452583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,5120,0.12014044655693902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,8192,0.07196178038914998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,5120,0.1916968954934014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,5120,0.3143848843044705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,4096,0.10059822267956203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,5120,0.050853331883748375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,6144,0.14395822419060603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,3584,0.08975466754701401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,4096,0.156622224383884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,4096,0.04202488727039761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,4096,0.32903822263081867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,3072,0.07693511247634888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,3584,0.03734666771358914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,3584,0.13613422711690268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,3584,0.3235537740919325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,3072,0.12092799610561794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,3072,0.03242044316397773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,2560,0.10408622026443481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,3072,0.4187902344597711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,2048,0.05347111158900791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,2560,0.3187733226352268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,2560,0.029278222057554457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,2048,0.02498666610982683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,2048,0.08583111233181423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,1536,0.04208088914553324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,2048,0.31845956378512913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,1536,0.019998222589492798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,1536,0.0682248870531718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,1024,0.029288000530666772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,1536,0.3282444477081299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,1024,0.050809777445263334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,1024,0.017050666941536795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,768,0.02296000056796604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,1024,0.32281332545810276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,768,0.04225777917438083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,512,0.017391110459963482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,768,0.01590666671593984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,768,0.3412906593746609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,512,0.033231112692091204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,512,0.3490320046742757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,512,0.014818666709793938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,256,0.013359111216333179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,2560,0.06816711028416951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,256,0.028952889972262915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,128,0.013462222284740873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,256,0.35394401020473903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,128,0.027464000715149775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,128,0.01348888874053955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,10240,128,0.3580906656053331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,64,0.011771555576059552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,10240,32,0.011830222275522021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,64,0.027313777142100867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,10240,32,0.02718488872051239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,10240,256,0.013608888619475894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,65536,1.279223124186198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,16384,0.2963555653889974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,65536,0.43527287907070583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,16384,0.5502239863077799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,65536,2.360690646701389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,16384,0.10621600018607245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,12288,0.2279155519273546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,16384,0.6324968867831761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,12288,0.4075208769904242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,12288,0.08283999893400404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,10240,0.18560977776845297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,12288,0.4829689131842719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,65536,2.374031914605035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,10240,0.0725039972199334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,8192,0.16255377398596868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,10240,0.33829421467251247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,10240,0.4102399879031711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,7168,0.13338755236731634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,8192,0.2745048999786377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,8192,0.34017332394917804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,8192,0.05916800101598104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,6144,0.11350133683946396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,7168,0.054177777634726636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,7168,0.23227556546529135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,7168,0.38071288002861875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,5120,0.09904978010389541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,6144,0.04847466614511278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,6144,0.19885867171817354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,6144,0.32449155383639866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,4096,0.08294844627380371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,5120,0.168412446975708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,5120,0.041972445117102734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,5120,0.3293200069003635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,3584,0.0696257750193278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,4096,0.1351111067665948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,4096,0.03437422381507026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,4096,0.3274471229977078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,3072,0.06290844413969252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,3584,0.3265724447038439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,3584,0.030730668041441176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,3072,0.10652533504698013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,3072,0.027449778384632532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,2560,0.05098399851057264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,3072,0.33166578080919057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,2560,0.09137866894404094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,2560,0.024902222885025874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,2560,0.3232399887508816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,2048,0.041639112763934664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,2048,0.07580355803171794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,2048,0.020834666159417894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,1536,0.03224710954560174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,2048,0.3239111105600993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,3584,0.12044444349077012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,1024,0.022804444034894306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,1536,0.01742755538887448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,1536,0.05905244747797648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,1536,0.32613688045077854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,1024,0.04415555463896858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,768,0.01936533384852939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,1024,0.014833778142929077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,1024,0.31274578306410045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,768,0.036355555057525635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,512,0.016571554872724745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,768,0.01385155568520228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,768,0.3220764531029595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,256,0.012667555775907306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,512,0.029333333174387615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,512,0.012859555582205454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,512,0.315611547893948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,256,0.026187555657492742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,256,0.012201777762836881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,128,0.024587555064095393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,8192,128,0.012141333685980903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,64,0.010737777584128909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,128,0.3749759991963704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,32,0.01075111081202825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,64,0.024420445164044697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,8192,32,0.024033778243594702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,8192,128,0.01016533292002148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,8192,256,0.3289253446790907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,65536,1.2079608705308702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,65536,0.3981413311428494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,16384,0.31139643987019855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,16384,0.5164248678419325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,16384,0.5925502247280544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,65536,2.170609792073568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,12288,0.18301155832078722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,65536,2.2102809482150607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,12288,0.3794168896145291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,12288,0.07406310902701484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,10240,0.15907911459604898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,12288,0.4648328887091742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,16384,0.09626489215426975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,10240,0.06583822435802884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,10240,0.31496887736850315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,10240,0.3920444382561578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,8192,0.13359555933210585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,8192,0.05375022358364529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,8192,0.3555742104848226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,7168,0.11617333359188503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,7168,0.215484446949429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,6144,0.09592444366878933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,7168,0.047820445564058095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,6144,0.18479377693600127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,6144,0.042597333590189614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,8192,0.25368089146084255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,6144,0.46342844433254665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,5120,0.15426577462090388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,7168,0.3640266789330377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,4096,0.0668026672469245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,5120,0.40505067507425946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,5120,0.0376506679587894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,4096,0.12623111406962076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,4096,0.030568000343110826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,3584,0.05743822124269274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,4096,0.3840720123714871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,3584,0.1110924416118198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,5120,0.08228177494472928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,3072,0.0485208895471361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,3584,0.34352800581190324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,3072,0.09961155388090347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,3072,0.02476799984773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,2560,0.0407626661989424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,3072,0.4280800024668376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,2560,0.08539999855889215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,2048,0.03333688775698344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,2560,0.32818577024671763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,2560,0.0225582222143809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,2048,0.06950310866038005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,1536,0.0262435558769438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,2048,0.019004444281260174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,2048,0.31894755363464355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,3584,0.027611555324660406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,1024,0.019719110594855416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,1536,0.05576088693406847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,1536,0.016512889001104567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,1024,0.04152266515625848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,1024,0.01425244410832723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,768,0.01499022212293413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,1024,0.33120089107089573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,768,0.035030222601360746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,512,0.01185066666867998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,768,0.013396444420019785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,768,0.34761065906948513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,512,0.02751200066672431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,256,0.009358222285906473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,512,0.011872000164455838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,512,0.3208746645185682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,256,0.023765333824687537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,128,0.008397333323955536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,256,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,256,0.3374657895829942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,128,0.023023110296991136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,64,0.007646222081449296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,1536,0.3869182268778483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,7168,128,0.011372444530328115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,7168,128,0.3041920132107205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,64,0.022978666755888198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,7168,32,0.022315555148654517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,65536,0.8706960148281522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,7168,32,0.007785778078767989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,16384,0.22535644637213814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,65536,0.38111376762390137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,16384,0.4685351053873698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,65536,1.9654818640814886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,16384,0.6197893354627821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,12288,0.17792088455624053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,16384,0.09069155322180854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,12288,0.07044533226225111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,12288,0.34880712297227645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,65536,2.180934270222982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,12288,0.4891360070970323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,10240,0.28983378410339355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,10240,0.06102489100562202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,8192,0.12088533242543538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,10240,0.38587021827697754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,8192,0.23235556814405653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,8192,0.049392888943354286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,7168,0.10497688584857517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,8192,0.3454079892900255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,7168,0.19999289512634277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,10240,0.15260888470543757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,7168,0.3028168943193224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,6144,0.0855226649178399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,6144,0.17073955800798205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,6144,0.3674000104268392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,5120,0.0761786699295044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,6144,0.04038577940728929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,5120,0.14335644245147705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,7168,0.04656622144911024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,5120,0.03600622216860453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,4096,0.06525688701205783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,4096,0.11614933278825547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,4096,0.38048532274034286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,4096,0.02865866820017497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,3584,0.05606222152709961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,3584,0.10325422551896836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,3584,0.36298667060004336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,3584,0.026055999928050574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,3072,0.049230221245023936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,5120,0.4572960005866156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,3072,0.08988089031643337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,2560,0.03525777657826742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,3072,0.024196444286240473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,3072,0.3516222106085883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,2560,0.0794373353322347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,2048,0.029001779026455347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,2560,0.4095822175343831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,2048,0.0642933315700955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,1536,0.023071999351183575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,2048,0.01789422167672051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,1536,0.0511084430747562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,1536,0.3998400105370416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,1536,0.015322667029168872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,1024,0.01905866629547543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,2560,0.02127200033929613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,1024,0.03852444556024339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,1024,0.013273778061072031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,768,0.01388088862101237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,1024,0.3860817750295003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,2048,0.3286497857835558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,512,0.013030221892727746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,768,0.012472888661755456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,768,0.3764657709333632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,512,0.026058667235904272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,256,0.010371555884679159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,512,0.01163999984661738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,256,0.02236266599761115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,256,0.011094222466150919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,256,0.3629768954383002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,128,0.010048000348938836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,128,0.021327111456129286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,768,0.03292266527811686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,64,0.009707555174827576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,6144,128,0.011018666956159802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,6144,32,0.00999111102686988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,64,0.021368889345063105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,6144,32,0.020991111795107525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,512,0.38338666492038304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,65536,0.8923724492390951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,6144,128,0.3065626621246338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,16384,0.19869511657290986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,65536,0.3417644500732422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,16384,0.4290764331817627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,65536,1.7957447899712458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,16384,0.08139555321799384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,12288,0.1505191061231825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,16384,0.5591688685946995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,12288,0.32234488593207467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,12288,0.06401866674423218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,10240,0.11082400215996636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,12288,0.44147822591993546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,10240,0.2679999934302436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,10240,0.05424088901943631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,8192,0.09609778059853448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,10240,0.3676328923967149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,8192,0.04486933350563049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,8192,0.21406932671864828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,8192,0.3370337751176622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,7168,0.18288800451490614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,7168,0.33708710140652126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,65536,2.051749335394965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,7168,0.040656887822681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,6144,0.07488177882300483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,6144,0.03686844309171041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,6144,0.15599466694725886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,6144,0.3283280001746284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,7168,0.08227022488911946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,5120,0.13095200061798096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,5120,0.33215199576483834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,4096,0.047446221113204956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,5120,0.03176800078815884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,4096,0.1062577764193217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,4096,0.025984888275464375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,3584,0.042671110894944936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,4096,0.4145048989189996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,3584,0.09435466925303142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,3072,0.03682488865322537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,3584,0.32706133524576825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,3584,0.023918223049905565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,3072,0.08147822486029731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,3072,0.021165332860416834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,3072,0.32590754826863605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,2560,0.06847466362847222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,2560,0.41952533192104763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,2560,0.019164444671736825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,2048,0.0258906665775511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,5120,0.06371821959813435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,2560,0.030543999539481267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,2048,0.0587715572781033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,1536,0.02008800043000115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,2048,0.34080712000528973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,2048,0.016215999921162922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,1536,0.04644799894756741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,1024,0.015015999476114908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,1536,0.013932444983058505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,1536,0.35334134101867676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,1024,0.03523555397987366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,768,0.01260355528857973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,1024,0.012270222107569376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,1024,0.3699822160932753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,768,0.02991022335158454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,768,0.011568889021873474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,512,0.010986666712496014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,768,0.3378906779819065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,512,0.023785778217845496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,512,0.011061333119869232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,512,0.37859466340806747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,256,0.020948444803555805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,256,0.010455999937322406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,128,0.008295999632941352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,256,0.37294310993618435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,128,0.019951999187469482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,64,0.007634667058785756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,5120,128,0.3034702142079671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,32,0.00795733349190818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,5120,128,0.010478222535716163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,64,0.01993777851263682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,5120,256,0.009104889300134447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,5120,32,0.019319999549123976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,65536,0.5672284232245551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,65536,0.28286756409539116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,16384,0.15518932872348362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,16384,0.38990399572584367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,16384,0.06928266419304742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,12288,0.11660799715254043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,65536,1.5944257312350805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,12288,0.29420532120598686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,12288,0.40439645449320477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,10240,0.10282933049731785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,12288,0.052729778819613986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,65536,1.915861341688368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,10240,0.046804444657431714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,10240,0.2439511087205675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,16384,0.5506088998582628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,10240,0.351839992735121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,8192,0.03829777903027005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,8192,0.19522755675845674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,7168,0.07163466347588433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,8192,0.36950045161777073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,7168,0.1668808857599894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,7168,0.035247110658221774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,7168,0.33440266715155703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,6144,0.06271111302905612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,6144,0.03128444486194187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,6144,0.1424648894204034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,5120,0.054899556769265064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,6144,0.32373778025309247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,8192,0.08307466904322307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,4096,0.041893333196640015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,5120,0.02730488942729102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,5120,0.12016710970136855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,5120,0.3364746570587158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,4096,0.09813511371612549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,3584,0.03851644529236688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,4096,0.022674666510687932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,4096,0.3104524347517225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,3072,0.032475554280810885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,3584,0.08715289168887669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,3584,0.02055111030737559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,3584,0.32920977804395885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,2560,0.028762665059831407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,3072,0.07422933313581678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,3072,0.018394667241308425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,3072,0.3216506640116374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,2560,0.0642791125509474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,2048,0.023704888092146978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,2560,0.016563556260532804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,2560,0.3270871109432644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,1536,0.018597332967652213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,2048,0.052413334449132286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,2048,0.014231999715169271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,2048,0.3168151113722059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,1024,0.013952889376216464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,1536,0.043097777499092944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,1536,0.013101333545313941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,1536,0.3356817828284369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,1024,0.031947556469175555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,768,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,1024,0.011520889070298938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,1024,0.2784319983588325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,512,0.008667555948098501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,768,0.010796444283591377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,768,0.3242551220787896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,512,0.021977777282396953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,512,0.010021333065297868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,256,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,512,0.44179288546244305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,256,0.019276445110638935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,256,0.009730666875839233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,128,0.006646222124497096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,256,0.42940711975097656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,768,0.027216888136333887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,128,0.018340445227093168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,4096,128,0.009671111073758867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,64,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,4096,32,0.0069644442862934535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,4096,128,0.37174754672580296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,64,0.01832444469134013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,4096,32,0.017873777283562552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,65536,0.5627617835998535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,65536,0.2380195458730062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,16384,0.15098044607374403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,65536,1.5144497553507488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,16384,0.38690932591756183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,12288,0.11211378044552273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,16384,0.05990311172273424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,16384,0.4869759877522786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,65536,1.7821049160427516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,10240,0.10142399867375691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,12288,0.04580444428655836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,12288,0.2798782189687093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,12288,0.43169599109225804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,10240,0.04079466727044847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,10240,0.2318506770663791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,10240,0.33436089091830784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,8192,0.033266666862699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,8192,0.18535378244188097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,7168,0.0731111102634006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,8192,0.353129784266154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,7168,0.029543999168607924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,7168,0.15865333875020346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,6144,0.06237600247065226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,7168,0.3301520082685682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,6144,0.026522666215896606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,6144,0.13556267155541316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,5120,0.05369777811898125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,6144,0.3523351086510553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,8192,0.08276266521877713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,4096,0.04374044471316867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,5120,0.023557333482636347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,5120,0.11461067199707031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,5120,0.32236001226637095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,4096,0.09331644243664211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,3584,0.03818044397566054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,4096,0.019846222466892667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,4096,0.3109644518958198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,3072,0.03233955634964837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,3584,0.017998221847746108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,3584,0.31174132559034556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,3072,0.07261244455973308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,2560,0.027143999934196472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,3072,0.4166595670912001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,2560,0.06104711029264662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,2560,0.014391111003028022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,2048,0.02387822171052297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,2560,0.4049164454142253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,2048,0.05025510986646017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,3584,0.08228533135520087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,1536,0.01866222255759769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,2048,0.012820444173283048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,1536,0.040807998842663236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,3072,0.016387555334303115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,1024,0.014324444863531323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,1536,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,1536,0.32310401068793404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,1024,0.03020533257060581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,768,0.011389333340856763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,1024,0.010294222169452244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,768,0.026114667455355327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,768,0.009718221922715506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,768,0.3256017896864149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,512,0.009150221943855286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,512,0.021261332763565913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,2048,0.3815111054314508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,256,0.0069431112044387394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,512,0.009362666971153682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,256,0.018345778187115986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,256,0.3262888855404324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,128,0.006242666807439592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,1024,0.32335466808742946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,128,0.01762488815519545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,128,0.3084159957038032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,128,0.00868088917599784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,64,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3584,32,0.00564177789621883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3584,512,0.33192266358269584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,64,0.01719555589887831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3584,32,0.01756266587310367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,65536,0.4211875597635905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3584,256,0.009007111191749573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,16384,0.11767733097076416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,65536,0.23476176791720918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,16384,0.3515155580308702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,65536,1.4324827194213867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,16384,0.5044053395589193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,12288,0.0899733304977417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,16384,0.05715733104281955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,12288,0.2658524513244629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,12288,0.3857777648501926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,10240,0.0734986662864685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,12288,0.04470488760206434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,10240,0.2201004425684611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,10240,0.4296106762356228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,8192,0.05973777506086561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,10240,0.03832533293300205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,65536,1.781201680501302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,8192,0.17627199490865073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,7168,0.05108977688683403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,8192,0.3744560082753499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,8192,0.03209422363175286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,7168,0.15120355288187662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,7168,0.029463112354278564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,6144,0.04300711221165127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,7168,0.35281509823269314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,6144,0.025595555702845257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,6144,0.1298275523715549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,5120,0.0380586650636461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,6144,0.3514346546596951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,4096,0.031934221585591636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,5120,0.0229751103454166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,5120,0.10849689112769233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,5120,0.34433333079020184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,4096,0.08961777554617988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,3584,0.02899644441074795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,4096,0.3137279881371392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,4096,0.01927911076280806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,3584,0.08014844523535834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,3072,0.025009777810838487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,3584,0.017613333132531907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,3584,0.3299146758185493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,2560,0.020970667401949566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,3072,0.06906400124231975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,3072,0.015634665886561077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,2560,0.05948888593249851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,2048,0.01707555519209968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,3072,0.4512542088826497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,2560,0.3207422097524007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,2048,0.047945777575174965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,2048,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,1536,0.013778666655222574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,2048,0.4191137684716119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,1536,0.0389048887623681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,1024,0.010570666856235929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,1536,0.3346071243286133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,1536,0.011438222395049201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,1024,0.02916711237695482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,768,0.008965333302815756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,1024,0.010545777777830759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,1024,0.3270622094472249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,768,0.025097777446111042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,512,0.008107555409272512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,768,0.009712888962692684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,512,0.020634666085243225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,768,0.4157591130998399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,256,0.0069164443347189165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,512,0.009054222040706212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,512,0.32882221539815265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,256,0.017652443713612027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,256,0.008650666309727563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,128,0.006676444576846228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,256,0.3396880096859402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,128,0.016920000314712524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,64,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,2560,0.014278221461508008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,3072,128,0.3827582200368245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,3072,32,0.006566222343179915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,64,0.016807110773192514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,3072,32,0.016584888100624084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,65536,0.36190043555365664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,65536,0.21946133507622612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,16384,0.10360088613298203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,3072,128,0.008735111190213097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,65536,1.3352844450208876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,16384,0.35896889368693036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,12288,0.079276442527771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,16384,0.4848657713996039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,16384,0.05719466341866387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,12288,0.04488977789878845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,12288,0.25383467144436306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,12288,0.37475021680196124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,10240,0.06552355819278292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,65536,1.7746729320949977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,10240,0.03853066762288412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,8192,0.05044800043106079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,10240,0.21012532711029053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,10240,0.31995201110839844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,7168,0.04534666736920675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,8192,0.030609776576360066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,8192,0.16609866089291042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,7168,0.14397688706715903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,7168,0.028374221589830186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,6144,0.03943199912707011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,7168,0.3291928768157959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,6144,0.12275200419955784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,6144,0.3557724422878689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,5120,0.03372622198528714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,6144,0.025240888198216755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,5120,0.1029813355869717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,5120,0.38646133740743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,5120,0.022135999467637803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,8192,0.3870533307393392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,4096,0.08515289094712998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,4096,0.01849333279662662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,4096,0.4181199868520101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,3584,0.0236462222205268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,3584,0.07522933350669013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,3584,0.36334845754835343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,3072,0.020964443683624268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,3584,0.016993777619467843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,3072,0.06485066811243693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,3072,0.01553066737122006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,3072,0.44126934475368923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,2560,0.018664888209766813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,2560,0.05551822317971123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,2560,0.36264623536003965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,2048,0.015601777368121676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,2560,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,4096,0.02696177860101064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,1536,0.012825777961148156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,2048,0.044828444719314575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,2048,0.012671999633312225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,1536,0.036714666419559054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,1024,0.008637333081828224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,1536,0.31879377365112305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,1536,0.011031111081441244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,1024,0.02778488728735182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,768,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,1024,0.010030222435792288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,1024,0.3207706610361735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,768,0.023991111252042983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,512,0.007103111180994246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,768,0.009424000150627559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,512,0.019658666517999437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,768,0.40433422724405926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,256,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,512,0.3259920014275445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,512,0.00904622260067198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,2048,0.4020506540934245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,256,0.01683822274208069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,128,0.0053013331360287136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,256,0.008685333033402761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,256,0.28030310736762154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,64,0.005236444373925527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2560,128,0.008648888932334052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2560,32,0.0052284445199701525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,64,0.01611288885275523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2560,128,0.3693315452999539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,32,0.015876443849669564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,65536,0.28566310140821666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,65536,0.17311378320058188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,16384,0.0837564468383789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2560,128,0.015970667203267414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,16384,0.3144444359673394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,65536,1.2456355624728732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,12288,0.06285333633422852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,16384,0.04523555437723795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,16384,0.4455368783738878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,12288,0.03538666831122504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,12288,0.2374497784508599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,65536,1.6420684390597875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,12288,0.34900177849663633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,10240,0.030421333180533514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,8192,0.0442737771405114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,10240,0.33393155203925234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,10240,0.19755733013153076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,8192,0.1564755572213067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,8192,0.025057777762413025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,7168,0.037626667155159846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,8192,0.3283697764078776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,7168,0.023392000132136877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,7168,0.1347928841908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,6144,0.03359377715322707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,7168,0.3329288959503174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,10240,0.04952088991800944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,5120,0.028063111835055884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,6144,0.020764445265134174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,6144,0.11458667119344075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,6144,0.3168577882978651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,4096,0.02384177843729655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,5120,0.09689244296815659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,5120,0.01883288886811998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,5120,0.3258355458577474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,3584,0.021106666988796655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,4096,0.015674667225943673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,4096,0.3048764334784614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,3584,0.07022933165232341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,3584,0.014510222607188754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,3584,0.4062995645734999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,3072,0.06033866935306125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,3072,0.01312444441848331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,2560,0.016193777322769165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,3072,0.3690097861819797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,2560,0.051550222767723926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,4096,0.07929599947399564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,2560,0.012078222301271228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,2048,0.013592888911565145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,3072,0.018584888842370775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,2048,0.04181866513358223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,2048,0.34713244438171387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,1536,0.011026666396194033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,2048,0.010682666467295753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,1536,0.03440355592303806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,1536,0.3905342155032688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,1536,0.009786666267448002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,1024,0.00886933339966668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,2560,0.43335644404093426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,1024,0.02614044480853611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,768,0.007762666377756331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,1024,0.0090515555606948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,768,0.022626666559113398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,768,0.008649778034951951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,512,0.006589333216349284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,768,0.4342302216423883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,512,0.01860000027550591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,512,0.008345777789751688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,256,0.005513777749405966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,512,0.4312568770514594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,1024,0.3599546750386556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,256,0.016227554943826463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,128,0.004926222066084544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,128,0.01551466683546702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,256,0.36150399843851727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,64,0.00461422238085005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,2048,128,0.28740088144938153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,128,0.007989333735571967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,2048,32,0.00490133340160052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,64,0.01516711049609714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,2048,32,0.015207999282413058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,65536,0.24978576766120064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,65536,0.169558220439487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,16384,0.0682613319820828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,2048,256,0.008085333638721043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,16384,0.30085423257615834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,65536,1.1696933110555012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,16384,0.044165333112080894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,12288,0.05363288852903578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,16384,0.44507646560668945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,12288,0.03418133325046963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,12288,0.22685334417555067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,10240,0.0436426666047838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,65536,1.6371688842773438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,10240,0.18819555971357557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,10240,0.029283556673261855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,10240,0.32131732834710014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,8192,0.034161776304244995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,8192,0.024784889486100938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,8192,0.14838933944702148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,7168,0.03035911255412632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,8192,0.31812000274658203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,7168,0.1282328897052341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,7168,0.022034666604465906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,6144,0.026896889011065166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,7168,0.33913421630859375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,6144,0.10927999681896633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,12288,0.40096176995171445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,6144,0.020710221595234342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,5120,0.022456000248591106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,6144,0.3259689013163249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,4096,0.019489778412712943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,5120,0.018227555685573153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,5120,0.09169689151975845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,5120,0.29514755143059623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,4096,0.07559200127919515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,4096,0.015544000599119397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,4096,0.32611287964714897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,3584,0.06645777490403917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,3072,0.015071110592948066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,3584,0.014095111025704278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,3072,0.056757330894470215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,3584,0.3923502233293321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,2560,0.012700444294346703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,3072,0.013136000268989138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,3072,0.3228684531317817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,2560,0.048691554201973804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,2560,0.011720889144473605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,2048,0.01094133324093289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,2560,0.32466665903727215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,3584,0.017300445172521803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,1536,0.009145778086450364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,2048,0.04019999835226271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,2048,0.010372444159454769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,1536,0.03291733397377862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,2048,0.39455376731024844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,1024,0.007589333587222629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,1536,0.009637333452701569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,1536,0.32006311416625977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,768,0.006608888920810487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,1024,0.009029332962301042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,1024,0.02476355599032508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,768,0.021368000242445204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,512,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,768,0.008715555899673039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,768,0.3120995627509223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,512,0.01772799922360314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,512,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,256,0.004953777624501122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,512,0.46117422315809464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,256,0.01517777807182736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,256,0.007992888490358988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,256,0.43662577205234104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,128,0.004557333472702238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,1024,0.32014399104648167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1536,128,0.007989333735571967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,128,0.01482311056719886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,64,0.004219555606444676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1536,32,0.00425155543618732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,64,0.0144533332851198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1536,32,0.014463111758232117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,65536,0.16770311196645102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1536,128,0.34048088391621906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,65536,1.0796107186211479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,16384,0.048453334305021495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,65536,0.12061156166924371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,16384,0.2792782253689236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,65536,1.499683592054579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,16384,0.4110115634070502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,12288,0.0355697770913442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,16384,0.03362844387690226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,10240,0.03022577696376377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,12288,0.21166043811374238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,12288,0.3942577838897705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,10240,0.17603645059797499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,8192,0.02297244469324748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,10240,0.022153778208626643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,10240,0.3244640032450358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,8192,0.13921599917941624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,8192,0.018498667412334018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,7168,0.020592000749376085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,8192,0.3763066662682428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,7168,0.11942044893900554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,6144,0.0184542222155465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,7168,0.3249617682562934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,7168,0.017313778400421143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,12288,0.026123555170165166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,6144,0.1020231114493476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,6144,0.01551999979548984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,6144,0.3128266599443224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,5120,0.0856319997045729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,4096,0.013145777914259167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,5120,0.40596622890896267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,4096,0.07019466824001737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,4096,0.012082666986518435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,3584,0.012335999972290464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,4096,0.3933928807576497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,3584,0.06153777572843763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,5120,0.016524444023768108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,3072,0.010950222611427307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,3584,0.011495999991893768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,3584,0.38794755935668945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,3072,0.05218933357132805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,5120,0.014560888210932413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,2560,0.00962311112218433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,3072,0.01071999967098236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,3072,0.3159484333462185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,2560,0.04498755600717333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,2048,0.008194666769769456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,2560,0.318869325849745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,2560,0.010046222143703038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,1536,0.007346666521496243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,2048,0.03683200147416856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,2048,0.009016888837019602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,2048,0.3110533290439182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,1024,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,1536,0.030035555362701416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,1536,0.008649778034951951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,1536,0.3216639889611138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,768,0.005590222362014983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,1024,0.008040000167157915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,768,0.020597333709398907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,1024,0.3875617716047499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,512,0.0047022220161226065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,768,0.007663110891977946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,768,0.324651559193929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,512,0.017293334007263184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,512,0.007320889168315464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,256,0.004230222354332606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,512,0.3267146746317546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,256,0.014452444182501899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,128,0.003914666672547658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,256,0.007287110719415877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,1024,0.02341777748531765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,256,0.39758843845791286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,64,0.0035537779331207275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,1024,128,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,1024,128,0.29792356491088867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,1024,32,0.0038888889054457345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,64,0.013542222479979197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,32,0.013697778185208639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,65536,0.1450773345099555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,65536,0.1192515558666653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,16384,0.0421982208887736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,1024,128,0.013798221945762634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,16384,0.26971377266777885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,65536,1.0393120447794597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,16384,0.41024798817104763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,16384,0.03314222229851617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,12288,0.03180799881617228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,65536,1.494424925910102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,12288,0.20490488741132948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,12288,0.0251857770813836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,12288,0.3830568790435791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,10240,0.027307555079460144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,10240,0.02204888893498315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,10240,0.16891822550031874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,8192,0.023158222436904907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,10240,0.37983910242716473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,8192,0.018221333622932434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,7168,0.02066044509410858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,8192,0.4461804495917426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,7168,0.1150311099158393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,7168,0.01686311099264357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,6144,0.01828533411026001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,7168,0.4419857660929362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,6144,0.09796888960732354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,8192,0.133788439962599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,6144,0.015640000502268474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,6144,0.41481600867377383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,5120,0.015607110328144498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,5120,0.014449778530332776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,5120,0.08239022228452894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,4096,0.012906666431162091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,5120,0.38561511039733887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,4096,0.06742844316694471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,4096,0.33713777860005695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,4096,0.0122417774465349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,3584,0.0587902201546563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,3072,0.010483555495738983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,3584,0.0118222220076455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,3584,0.3253493309020996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,3072,0.050501333342658154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,2560,0.009418666362762451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,3072,0.010762666662534079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,2560,0.04353777898682488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,3072,0.40257244639926487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,2048,0.008257778154479133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,2560,0.010033778018421596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,2560,0.32578221956888836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,3584,0.012147555748621622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,1536,0.007381333245171442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,2048,0.03620355659061008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,2048,0.009329777624871995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,2048,0.31927556461758083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,1024,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,1536,0.029013332393434312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,1536,0.008646222452322641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,1536,0.35993154843648273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,1024,0.008021333151393467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,1024,0.022672888305452134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,768,0.005543111099137201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,768,0.019911999503771465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,768,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,512,0.004921777794758479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,768,0.3205946551428901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,512,0.0162471118900511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,256,0.004625777817434735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,512,0.007330666813585494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,512,0.33418843481275773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,256,0.013956444131003486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,128,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,256,0.007246221933099959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,256,0.32261421945359975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,128,0.013335111240545908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,64,0.0038808890514903595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,128,0.30519644419352215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,768,32,0.00387999994887246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,64,0.013134222063753339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,768,1024,0.3791128794352214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,65536,0.09218044413460626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,768,32,0.013074666261672974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,65536,0.11812267038557266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,16384,0.03373600045839945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,16384,0.25993598832024467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,768,128,0.0069857777820693115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,65536,0.9984169006347656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,12288,0.024205333656734888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,16384,0.032849778731664024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,16384,0.4066755506727431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,12288,0.19712532891167533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,12288,0.3255857891506619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,65536,1.4762996037801106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,10240,0.021171554923057556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,12288,0.02566577825281355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,8192,0.017238222890430026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,10240,0.021130666136741638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,10240,0.1631457805633545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,10240,0.31384889284769696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,8192,0.1289564503563775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,7168,0.015040889382362366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,8192,0.01774488886197408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,7168,0.11093066798316108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,8192,0.3896328873104519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,6144,0.0131217779384719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,7168,0.3061457739935981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,6144,0.09368444151348537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,6144,0.3700515429178874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,5120,0.011492444409264458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,6144,0.015409777561823526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,5120,0.07964266671074761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,5120,0.014157333307796054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,5120,0.4287288983662923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,4096,0.010447999669445885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,7168,0.016529776983790927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,4096,0.012115555504957834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,4096,0.06445866823196411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,3584,0.009313777916961247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,4096,0.3641457822587755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,3584,0.056187556849585645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,3584,0.011413333316644033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,3072,0.008377778033415476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,3584,0.3644115659925673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,3072,0.0483822226524353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,3072,0.01075644459989336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,2560,0.007610666255156199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,3072,0.40385066138373477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,2560,0.04176266656981575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,2048,0.006903999795516332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,2560,0.30340533786349827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,2560,0.009738666315873465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,2048,0.033560888634787664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,1536,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,2048,0.009104000197516548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,2048,0.31567467583550346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,1536,0.02780888809098138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,1024,0.005058666484223472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,1536,0.008731555607583787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,1024,0.022032888399230108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,1536,0.39583643277486164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,768,0.004600000050332812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,1024,0.008013333711359236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,768,0.019308444526460435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,512,0.004001777619123459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,768,0.3173066775004069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,768,0.007692444655630324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,512,0.016199999385409884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,256,0.0037208890749348533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,512,0.007352888584136963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,512,0.317467557059394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,256,0.013859555953078799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,256,0.006976000136799282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,128,0.003556444413132138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,256,0.38873688379923504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,128,0.013189333180586496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,64,0.0032017777363459268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,1024,0.30842222107781303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,512,128,0.007009777757856581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,512,32,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,512,128,0.3600640032026503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,64,0.012854222622182635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,65536,0.05908799833721585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,512,32,0.012535110943847232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,16384,0.01889511115021176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,65536,0.11767733097076416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,16384,0.24839554892645943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,16384,0.40610043207804364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,12288,0.01757866640885671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,16384,0.03124177787039015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,65536,0.9547359678480359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,12288,0.02344888945420583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,12288,0.188880893919203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,12288,0.34547821680704754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,65536,1.474013328552246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,10240,0.15684799353281656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,8192,0.012863111164834766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,10240,0.020959110723601446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,8192,0.12339644961886936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,8192,0.33102311028374565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,7168,0.011423110961914062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,8192,0.017895110779338412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,7168,0.10581066873338486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,10240,0.01536622146765391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,6144,0.010485333700974783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,7168,0.3344239923689101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,7168,0.01649599936273363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,6144,0.015311110350820752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,10240,0.33719733026292587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,5120,0.009111111362775167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,6144,0.32778310775756836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,5120,0.0758719974093967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,5120,0.014104889498816596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,4096,0.007996444072988298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,5120,0.3284897804260254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,4096,0.06163377894295586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,4096,0.012112889024946423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,3584,0.007649777664078607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,4096,0.31269510587056476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,3584,0.05340266558859083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,3584,0.011368888947698804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,3072,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,3584,0.33054934607611763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,3072,0.045960889922248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,6144,0.08957688675986396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,2560,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,3072,0.01070666644308302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,3072,0.33321600490146214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,2560,0.038626667526033186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,2048,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,2560,0.009752888646390703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,2560,0.33728978368971085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,2048,0.008975110948085785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,2048,0.03233777814441257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,1536,0.005249777601824866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,2048,0.36319200197855633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,1536,0.026914666096369427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,1536,0.008343111309740279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,1024,0.004249777644872665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,1536,0.3607190979851617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,1024,0.021324444148275588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,768,0.003900444342030419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,1024,0.007975111405054728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,1024,0.40255199538336855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,768,0.018595554762416415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,512,0.003847111016511917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,768,0.35321421093410915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,768,0.007688889073001013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,512,0.015517334143320719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,256,0.0035546666218174826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,512,0.007308444215191736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,512,0.32713243696424693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,256,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,128,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,256,0.006948444578382704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,128,0.012401777837011548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,256,0.41593511899312335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,64,0.0029253332565228143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,256,128,0.0070159998204973005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,256,32,0.0031439999325407874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,256,128,0.3171822230021159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,65536,0.049145778020222984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,64,0.012465777496496836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,256,32,0.012444444828563266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,65536,0.11742311053805882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,16384,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,16384,0.24778400527106392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,16384,0.4067404535081651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,65536,0.9536346859402127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,12288,0.010732444624106089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,16384,0.030763556559880573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,12288,0.1883315510219998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,12288,0.023436443673239812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,12288,0.3324311044481066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,65536,1.4745528962877061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,10240,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,8192,0.012454222473833295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,10240,0.020869332883093093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,10240,0.15699377324846056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,10240,0.31621066729227704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,8192,0.12227466371324326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,7168,0.01128000020980835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,8192,0.017878222796652053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,7168,0.10538578033447266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,8192,0.387884431415134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,6144,0.010069333016872406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,7168,0.31821245617336696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,6144,0.015160888433456421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,5120,0.009000000026490953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,6144,0.08945333295398289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,6144,0.33766667048136395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,5120,0.0754951106177436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,5120,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,5120,0.3549511167738173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,4096,0.060996446344587535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,4096,0.012054222325483957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,3584,0.007263110743628607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,7168,0.016537777251667447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,3584,0.052480889691246875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,3584,0.3605768945482042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,3584,0.011074666347768573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,3072,0.006646222124497096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,4096,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,3072,0.04439289040035672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,2560,0.005994666781690385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,3072,0.010455110834704505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,3072,0.3424657715691461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,2560,0.03772977656788296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,2048,0.005578666511509154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,2560,0.009716444545321995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,4096,0.3750622272491455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,2560,0.32587377230326336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,1536,0.004902222090297275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,2048,0.031874666611353554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,2048,0.0090506664580769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,1536,0.02679644525051117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,2048,0.32603110207451713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,1024,0.004234666625658671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,1536,0.00847999999920527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,1536,0.3228995535108778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,1024,0.02128711177243127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,768,0.0038720000949170855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,1024,0.008020444048775567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,1024,0.31455111503601074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,512,0.0038488888078265717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,768,0.018624888526068795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,768,0.007678222325113084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,768,0.31289956304762095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,512,0.015487111277050443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,256,0.0032151111712058387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,512,0.007302222152551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,512,0.3637244436475966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,256,0.012798222402731577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,128,0.0032133333798911837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,256,0.0069439998931354946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,256,0.3582995467715793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,128,0.012116444607575735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,64,0.003154666680428717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2048,128,128,0.0069253332912921906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,128,32,0.0032479998966058097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,64,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,65536,0.046275556087493896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,128,32,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,16384,0.010743111371994019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,16384,0.24794133504231772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,12288,0.009408888717492422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,65536,0.9540470971001519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,10240,0.011001778145631155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,12288,0.1877795590294732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,8192,0.009671111073758867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,10240,0.1562755505243937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,7168,0.009360888765917884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,8192,0.12187111377716064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,6144,0.00869955536392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,7168,0.10529955228169759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,5120,0.008285333712895712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,6144,0.08935199843512641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,4096,0.007642666498819987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,3584,0.007108444141017065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,5120,0.0753484434551663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,3072,0.006627555522653792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,4096,0.060885330041249595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,2560,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,3072,0.044608000251981944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,2048,0.005607999861240387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,3584,0.0524524450302124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,2560,0.03768710957633124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,1536,0.004974222017659081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,1024,0.004323555363549126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,2048,0.03204444381925795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,768,0.0039057777159743835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,1536,0.026397332549095154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,512,0.003563555578390757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,768,0.018485332528750103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,1024,0.021359110871950786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,256,0.0032497776879204642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,512,0.015134221977657743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,128,0.0031644445326593188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,64,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,128,0.012335999972290464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,256,0.012895111408498553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,64,32,0.0028880000528362063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,64,0.011953777737087674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,65536,0.04593511091338264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,16384,0.009602666729026372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,64,32,0.011790222591824003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,12288,0.008765333228641087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,16384,0.24774487813313803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,10240,0.00796622203456031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,12288,0.1882044474283854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,8192,0.012518222133318583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,10240,0.15598932902018228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,65536,0.9536488850911459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,7168,0.011386666860845355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,6144,0.01035733355416192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,8192,0.12219644917382134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,7168,0.10540444321102566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,5120,0.009701333112186855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,4096,0.008757332960764566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,6144,0.08929244677225749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,3584,0.008290666672918532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,5120,0.07522755861282349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,3072,0.006599999964237213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,4096,0.06071199973424276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,3584,0.0521342224544949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,2560,0.005984000033802456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,2048,0.005325333525737126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,3072,0.04420533445146349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,1536,0.004570666700601578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,2560,0.03763377666473389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,1024,0.0038702223036024305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,2048,0.032144000132878624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,1536,0.02679377794265747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,768,0.0038853333228164246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,1024,0.020941333638297185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,512,0.0032319999817344877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,256,0.003301333429084884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,768,0.01830222209294637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,128,0.0028773333049482773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,512,0.015160888433456421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,64,0.0027706666539112725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,256,0.012878222597969903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,128,0.012108444339699216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,64,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2048,32,32,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,16384,1.2775857713487413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2048,32,32,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,16384,1.4684524536132812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,16384,0.36904001235961914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,12288,1.067721790737576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,16384,1.5540346569485133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,12288,1.0507422553168404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,12288,0.2805546654595269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,10240,0.7441520161098905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,10240,0.8510239919026693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,12288,1.055586708916558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,10240,0.23832355605231392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,8192,0.67794402440389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,10240,0.9230986701117622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,8192,0.7172382142808703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,8192,0.1989004479514228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,7168,0.4862746662563748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,8192,0.6640835338168675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,7168,0.5988666746351454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,6144,0.42364711231655544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,7168,0.5715849134657118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,6144,0.5098986625671387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,7168,0.17054222689734566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,6144,0.15576088428497314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,6144,0.5057751337687174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,5120,0.3433537748124864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,5120,0.42592265870836044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,5120,0.13377155197991267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,4096,0.2812088860405816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,5120,0.4657760196261936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,4096,0.33988266521030003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,3584,0.24079733424716523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,4096,0.11202843983968098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,4096,0.3687893284691705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,3584,0.31898487938774955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,3072,0.21114133463965523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,3584,0.4379288885328505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,3072,0.2554977734883626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,3072,0.08567288849088882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,3072,0.4111013412475586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,2560,0.16411999861399332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,3584,0.09923466708925034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,2560,0.22321955362955728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,2560,0.07974221971299914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,2048,0.14608444107903376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,2560,0.3334088855319553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,2048,0.1832480033238729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,1536,0.11364444096883138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,2048,0.06688088840908475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,2048,0.38456087642245823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,1536,0.14643911520640054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,1536,0.05470844440990024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,1024,0.07526577843560113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,1536,0.3346506754557292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,1024,0.11053866810268825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,1024,0.043987555636299976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,768,0.062499556276533336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,1024,0.3398391140831842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,768,0.08942044443554348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,768,0.04058488872316148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,768,0.39299999343024355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,512,0.07795288827684191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,512,0.03643111056751675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,256,0.035339554150899254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,512,0.4047768910725911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,256,0.07132444116804335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,256,0.3995350996653239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,256,0.03354488809903463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,512,0.04683822062280443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,128,0.0296088887585534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,128,0.06673600276311238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,64,0.02678933408525255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,65536,128,0.03304800060060289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,65536,128,0.30949332978990346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,65536,32,0.025809778107537165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,64,0.06657866636912028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,65536,32,0.06649333238601685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,65536,1.186430189344618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,65536,1.7632204691569011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,65536,0.4085493352678087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,16384,0.2885013421376546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,16384,0.4256088998582628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,16384,0.1060746643278334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,16384,0.4677591323852539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,12288,0.2226355605655246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,65536,1.7905635833740234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,12288,0.31542756822374135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,12288,0.08080266581641303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,10240,0.18807021776835123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,12288,0.4064142174190945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,10240,0.26034045219421387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,10240,0.07111110952165392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,8192,0.14820978376600477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,10240,0.3526453442043728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,8192,0.05853511227501763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,8192,0.2144942283630371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,7168,0.13729243808322483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,8192,0.34894755151536727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,7168,0.1888959937625461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,7168,0.0518506666024526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,2048,128,128,0.3130719926622179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,6144,0.12148977650536431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,7168,0.3305155436197917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,6144,0.16104443868001303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,6144,0.04731377628114489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,5120,0.0966942244105869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,6144,0.3694302241007487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,5120,0.13896444108751085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,5120,0.04184177849027845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,4096,0.08080089092254639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,5120,0.37468356556362575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,4096,0.11459288332197402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,4096,0.03361777795685662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,3584,0.075217776828342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,4096,0.3834373421139187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,3584,0.02996444371011522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,3584,0.09930488798353408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,3072,0.06220799684524536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,3584,0.39020265473259824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,3072,0.08860355615615845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,3072,0.02720355490843455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,2560,0.053716444306903414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,3072,0.35228623284233945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,2560,0.07829066779878405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,2048,0.04792444573508369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,2560,0.024216888679398432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,2560,0.36345601081848145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,2048,0.020563556088341605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,1536,0.03650310966703627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,2048,0.36980356110466855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,1536,0.05043288734224108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,1536,0.346343994140625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,1536,0.01643111142847273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,1024,0.025477333201302424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,1024,0.0392159985171424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,1024,0.01440355512830946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,768,0.01795644395881229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,1024,0.33993866708543563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,768,0.032891554964913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,2048,0.06447022491031222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,768,0.013082666529549493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,512,0.015464888678656684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,768,0.3310871124267578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,256,0.012329777909649743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,512,0.028110222684012517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,256,0.025103110406133864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,512,0.3274879985385471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,128,0.010060444474220276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,256,0.011331555744012197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,256,0.32997600237528485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,128,0.02420888841152191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,64,0.010411555568377176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,128,0.011433777709801992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,16384,128,0.2970053354899089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,16384,32,0.010776888993051318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,64,0.023694222172101338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,16384,32,0.023972445064120825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,16384,512,0.012170666621790992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,65536,0.8471182187398275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,16384,0.28204443719651967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,65536,0.3514275550842285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,65536,1.4244213104248047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,16384,0.3876195483737522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,16384,0.08624000019497341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,12288,0.18129866653018525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,16384,0.4433022340138753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,12288,0.2553226682874891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,12288,0.06738399797015719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,10240,0.1379786729812622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,12288,0.41967021094428164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,10240,0.21516533692677817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,65536,1.5496657689412434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,10240,0.0607208874490526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,8192,0.12043555577596028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,10240,0.3270168834262424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,7168,0.10263466835021973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,8192,0.04895288745562235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,8192,0.32037244902716744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,7168,0.15265866120656332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,7168,0.04559377829233805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,6144,0.08719733026292588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,7168,0.4063395659128825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,6144,0.13513155778249106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,6144,0.03926488757133484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,6144,0.3321475452846951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,5120,0.1141973336537679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,8192,0.17890755335489908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,5120,0.3278648853302002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,4096,0.05775022506713867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,5120,0.034120000070995755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,4096,0.09438666370179917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,3584,0.04823733369509379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,4096,0.028259555498758953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,4096,0.326200008392334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,3584,0.08299555381139119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,3072,0.04670222269164192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,5120,0.0738240016831292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,3584,0.40267377429538304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,3072,0.074562668800354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,2560,0.03793599870469835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,3072,0.3032613330417209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,3072,0.02278577784697215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,2560,0.06403822369045682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,2560,0.3393875492943658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,2048,0.033471110794279314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,2560,0.020353777541054618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,3584,0.025095999240875244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,2048,0.053117334842681885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,1536,0.024873778223991394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,2048,0.3553848796420627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,2048,0.017826666434605915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,1536,0.04327466752794054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,1024,0.0170462214284473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,1536,0.014639111028777229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,1536,0.34147198994954425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,1024,0.03357777661747403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,768,0.014486221803559197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,1024,0.012532444463835822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,1024,0.3583368990156386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,768,0.028311110205120508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,768,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,512,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,768,0.3290097713470459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,512,0.023864888482623633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,256,0.010487111078368293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,512,0.010668444136778513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,512,0.368505769305759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,256,0.02165511084927453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,128,0.010335110955768162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,256,0.01036444471942054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,256,0.3398408889770508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,128,0.021399999658266704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,64,0.009695111049546136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,12288,128,0.0103502223889033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,12288,32,0.009617778162161509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,64,0.021015110943052504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,12288,32,0.020598222812016804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,65536,0.7407297558254666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,12288,128,0.31445421112908256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,65536,1.2637946870591905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,16384,0.19821155071258545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,65536,0.2999768786960178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,16384,0.3084799978468153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,12288,0.14987466070387098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,16384,0.3916337754991319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,12288,0.2271360026465522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,65536,1.3950764338175456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,12288,0.33372977044847274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,12288,0.060439109802246094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,10240,0.1268542210261027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,10240,0.19191200203365752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,10240,0.051544000705083214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,8192,0.09882399770948623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,10240,0.33201244142320424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,16384,0.07737333244747586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,8192,0.15460177262624106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,7168,0.08594666586981879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,8192,0.04309244288338555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,8192,0.3238817850748698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,6144,0.07521510786480375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,7168,0.03919555412398444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,7168,0.1349440018335978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,7168,0.37279200553894043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,6144,0.12027822600470649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,6144,0.03391022152370877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,6144,0.3284791045718723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,5120,0.1022613313462999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,5120,0.02961955467859904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,4096,0.05429244372579786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,5120,0.33808088302612305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,4096,0.08431555827458699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,4096,0.024752888414594863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,3584,0.0484479996893141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,4096,0.32983554734124076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,3584,0.07431199815538195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,3584,0.023043556345833674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,3072,0.03939377930429246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,3584,0.32631643613179523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,5120,0.062216891182793506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,3072,0.0665466652976142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,2560,0.03303022185961405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,3072,0.020222221811612446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,3072,0.3255066606733534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,2048,0.027543110979927912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,2560,0.01869155632125007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,2048,0.04629155662324694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,2560,0.3877982298533122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,1536,0.02256711158487532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,2048,0.32228533426920575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,2048,0.01587377819750044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,1536,0.03879377908176846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,1536,0.013166222307417126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,1024,0.016315556234783597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,1536,0.32065510749816895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,1024,0.029527111185921565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,1024,0.011446221835083432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,768,0.01367022262679206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,1024,0.3802951176961263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,2560,0.05759288867314657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,512,0.010880889164076911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,768,0.3280408912234836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,768,0.010746666954623329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,512,0.022051556242836848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,512,0.010112000008424124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,512,0.35887911584642196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,256,0.008339555727110969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,256,0.01997155613369412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,256,0.009664889011118147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,256,0.39467112223307294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,128,0.008250666989220513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,768,0.025480889611774023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,128,0.019711999429596793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,64,0.007620444728268518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,10240,128,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,10240,32,0.007616889145639207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,10240,128,0.339430226220025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,64,0.0195795562532213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,10240,32,0.019296889503796894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,65536,0.5530168745252821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,65536,0.24543645646837023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,65536,1.096015082465278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,16384,0.1525039937761095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,16384,0.2681271235148112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,12288,0.11358933978610569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,16384,0.06560977962281969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,65536,1.2663022147284615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,12288,0.1997164487838745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,10240,0.099144889248742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,12288,0.05110399921735128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,12288,0.5326693322923448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,10240,0.1665279997719659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,8192,0.08092800113889906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,10240,0.33262133598327637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,10240,0.04534399840566847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,16384,0.3509342140621609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,8192,0.13503289222717285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,7168,0.07084088855319552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,8192,0.036955555280049644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,7168,0.12089510758717854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,7168,0.03274666600757175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,6144,0.06003466579649183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,7168,0.3294435607062446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,6144,0.028910222980711196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,5120,0.05309066507551405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,5120,0.08952799770567153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,8192,0.3230728838178846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,5120,0.025236444340811834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,5120,0.3275768756866455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,4096,0.04452444447411431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,4096,0.07232177919811673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,6144,0.10522133111953735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,4096,0.3278702100118001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,3584,0.03901777664820353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,4096,0.020992888344658747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,3584,0.019350222415394254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,3072,0.033590220742755465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,3584,0.3831102318233914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,3072,0.05815200010935465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,3072,0.017435555656750996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,2560,0.02923111120859782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,3072,0.3837333255343967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,2560,0.05013333426581489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,6144,0.33901333808898926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,2048,0.023637332850032385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,2560,0.36972443262736004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,2560,0.015985776980717976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,3584,0.06622044245402019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,1536,0.019101333287027147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,2048,0.0417795545525021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,2048,0.014095111025704278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,1536,0.035192890299691096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,2048,0.36437421374850804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,1024,0.014155555102560254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,1536,0.31254222657945424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,1024,0.0269795556863149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,1024,0.010423111418883005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,768,0.0118568887313207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,1024,0.32272089852227104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,768,0.023520888553725347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,768,0.009713778065310584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,512,0.009757333331637913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,768,0.38984354337056476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,512,0.02003466586271922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,256,0.009302222066455418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,512,0.32248711585998535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,512,0.009382222261693742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,256,0.018622222873899672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,128,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,1536,0.012069332930776807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,256,0.008972444468074376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,256,0.309534231821696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,128,0.018245332770877414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,64,0.007667555577225155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,8192,32,0.007666666474607255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,8192,128,0.00870577742656072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,64,0.018262222409248352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,8192,32,0.017986666825082567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,65536,0.44206399387783474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,65536,0.20053954919179282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,65536,1.0247706307305229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,16384,0.1256542205810547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,16384,0.25295289357503253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,16384,0.4808933469984267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,65536,1.135306676228841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,12288,0.09873244497511123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,16384,0.05452711052364773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,8192,128,0.2808515495724148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,12288,0.18583556016286215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,10240,0.0840488870938619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,12288,0.042372445265452065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,12288,0.4452524185180664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,10240,0.1551440027025011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,10240,0.03757244348526001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,8192,0.06780888636906941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,10240,0.3623502254486084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,8192,0.126183999909295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,7168,0.059883554776509605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,8192,0.40268977483113605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,7168,0.11132621765136719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,7168,0.3515617847442627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,6144,0.058111998769972056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,7168,0.02719377809100681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,6144,0.09879644711812337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,8192,0.03012711140844557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,6144,0.42981865670945907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,6144,0.02479022244612376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,5120,0.05029866761631436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,5120,0.021385777327749465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,5120,0.08446311288409764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,4096,0.03955644369125366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,5120,0.34949154324001735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,4096,0.01898755629857381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,4096,0.06714844703674316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,3584,0.03636711173587375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,3584,0.05952088700400459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,3072,0.03188088867399428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,3584,0.016911110944218107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,3584,0.4032453431023492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,3072,0.0537395543522305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,2560,0.027112889620992873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,3072,0.015399111641777886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,3072,0.39687289132012266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,2560,0.0470142232047187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,2048,0.022256000174416438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,2560,0.0140364451540841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,2560,0.38324443499247235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,2048,0.03911822372012668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,4096,0.33637687895033097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,1536,0.017979555659823947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,2048,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,1536,0.032518221272362605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,2048,0.3570088810390896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,1024,0.012549333274364471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,1536,0.01075466639465756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,1536,0.3281537691752116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,768,0.0110880004035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,1024,0.02474222249454922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,1024,0.009452444811662039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,1024,0.32240088780721027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,512,0.008379555410808986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,768,0.009010666774378883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,768,0.34644177224900985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,512,0.018911111685964797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,256,0.00630044440428416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,512,0.008352888955010308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,256,0.017887110511461895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,512,0.39438666237725156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,128,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,256,0.33221154742770725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,256,0.007992888490358988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,128,0.017521777086787753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,7168,128,0.008044444852405125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,64,0.005227555417352253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,7168,32,0.005610666755172942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,7168,128,0.31075912051730686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,768,0.022311111291249592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,64,0.017080000705189176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,7168,32,0.016960889101028442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,65536,0.42847111490037704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,16384,0.13295200135972765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,65536,0.19594132900238037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,65536,0.9448391596476237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,16384,0.2332435448964437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,12288,0.0865795546107822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,16384,0.37721954451666934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,16384,0.0537120004494985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,12288,0.04246933261553446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,12288,0.17303910520341662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,10240,0.07141244411468506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,12288,0.36757421493530273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,10240,0.1454088952806261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,8192,0.05426844292216831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,10240,0.035767998960283064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,10240,0.3360559940338135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,8192,0.11913511488172744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,8192,0.029304001066419814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,7168,0.04861244559288025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,8192,0.3929137653774685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,7168,0.10646488931443955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,6144,0.04746133420202467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,7168,0.02652888827853733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,7168,0.32622488339742023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,65536,1.123172442118327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,6144,0.09330577982796563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,5120,0.04049066702524821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,6144,0.023429332507981196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,6144,0.41375290022955996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,5120,0.0803884466489156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,4096,0.03331911232736375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,5120,0.021166221963034734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,5120,0.35609155231051975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,4096,0.0637706650627984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,3584,0.029346668057971533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,4096,0.017708443933063082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,4096,0.43078843752543133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,3584,0.05645689037111071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,3072,0.023268444670571223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,3584,0.01640533407529195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,3584,0.3646506733364529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,3072,0.05000177688068814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,3072,0.014858666393491956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,2560,0.021261332763565913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,3072,0.38584176699320477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,2560,0.0446284446451399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,2048,0.01829600003030565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,2560,0.34868799315558535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,2048,0.03684355484114753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,2048,0.01186044431395001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,2048,0.3745528856913249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,1536,0.015247111519177755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,1536,0.031288888719346784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,1536,0.35396801100836855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,1536,0.010761777559916178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,1024,0.011595555477672152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,1024,0.023754666248957317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,2560,0.013958222336239286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,768,0.009300444689061906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,1024,0.3327813413408068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,1024,0.009406222237481011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,512,0.007846222155623967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,768,0.02095288866096073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,768,0.008646222452322641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,768,0.3263884385426839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,256,0.006674666785531574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,512,0.01834488908449809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,256,0.01717422240310245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,512,0.40330666965908474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,128,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,256,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,256,0.4142462147606744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,128,0.016864889197879367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,64,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,128,0.008033778104517195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,6144,128,0.2994115617540148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,6144,32,0.006409777949253718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,64,0.016556445095274184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,6144,32,0.016529776983790927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,65536,0.37123910586039227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,6144,512,0.00832266691658232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,16384,0.10262755552927653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,65536,0.19281421767340767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,16384,0.21221421824561226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,65536,0.8627066612243652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,12288,0.07849422428343031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,16384,0.34670400619506836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,12288,0.15782754951053196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,65536,1.1226897769504123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,12288,0.3373982111612956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,12288,0.04049866729312473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,10240,0.13364622328016493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,10240,0.03390044305059645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,10240,0.41631023089090985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,16384,0.05060088965627882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,8192,0.047170665529039174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,8192,0.02834577692879571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,8192,0.1093920005692376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,7168,0.040807998842663236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,8192,0.33796710438198513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,7168,0.09741777843899196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,7168,0.025044444534513686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,10240,0.06513333320617676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,6144,0.03851644529236688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,7168,0.3334177864922418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,5120,0.032095111078686185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,6144,0.08477777904934353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,6144,0.023056889573733013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,6144,0.3152915636698405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,4096,0.02815999918513828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,5120,0.07227022118038602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,5120,0.020640888147883944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,5120,0.31570400132073295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,4096,0.05806044737497965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,4096,0.017671111557218764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,4096,0.34163912137349445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,3584,0.051685333251953125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,3072,0.022413333257039387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,3584,0.016148444679048326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,3584,0.40250221888224286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,3072,0.046210666497548424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,2560,0.019169777631759644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,3072,0.014711111783981323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,3072,0.3376764456431071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,2560,0.040806220637427434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,2560,0.013233777549531726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,3584,0.02442755632930332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,2048,0.015953777564896476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,2560,0.36769866943359375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,2048,0.03414311011632284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,1536,0.012823111481136747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,2048,0.01182577759027481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,2048,0.35477156109280056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,1536,0.028071999549865723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,1024,0.009399111072222391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,1536,0.010091555615266165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,1536,0.3876239988538954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,1024,0.02200444373819563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,768,0.008329778081840938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,1024,0.009104000197516548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,1024,0.36436798837449813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,768,0.019595555133289762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,512,0.00700800038046307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,768,0.008656000097592672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,512,0.017279111676745944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,768,0.3932782279120551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,256,0.0063440000845326324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,512,0.3477164374457465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,512,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,256,0.01620444489849938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,256,0.007819555699825287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,128,0.005666666560702854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,256,0.3455839951833089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,128,0.015837333268589444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,64,0.005244444641802046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,5120,128,0.008025777836640676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,5120,32,0.005280888742870755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,5120,128,0.3835013442569309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,64,0.015569777952300178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,5120,32,0.015480000111791821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,65536,0.2441493405236138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,16384,0.06774044699139066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,65536,0.1496515538957384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,65536,0.7797244389851888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,16384,0.1947164403067695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,16384,0.039806223577923246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,12288,0.05292888813548618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,16384,0.40611643261379665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,12288,0.14271467261844212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,65536,0.9903173446655273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,12288,0.031015111340416804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,10240,0.04404266675313314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,12288,0.417980432510376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,8192,0.03819200065400865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,10240,0.4071982171800401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,8192,0.09891021913952297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,8192,0.022590221630202398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,7168,0.03526844580968221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,8192,0.40714044041103786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,7168,0.08843466970655654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,7168,0.020776889390415616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,6144,0.030544890297783747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,6144,0.07669244209925334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,6144,0.3781769010755751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,6144,0.018726223044925265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,5120,0.026177777184380427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,10240,0.12027643786536323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,5120,0.06473510795169406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,5120,0.33867377705044216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,5120,0.016916443904240925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,4096,0.022824888428052265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,4096,0.05207733313242594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,4096,0.33217332098219127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,3584,0.020488889680968392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,10240,0.026575111680560645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,3584,0.04731466703944736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,3584,0.013482666677898832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,3072,0.018026666508780587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,3584,0.34496355056762695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,3072,0.04332800043953789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,3072,0.33671289020114475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,2560,0.015306666493415833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,4096,0.014327110515700446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,2560,0.037090665764278836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,7168,0.3926515579223633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,2048,0.013939556148317126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,2560,0.31215911441379124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,2560,0.010854221880435944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,2048,0.031657778554492526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,2048,0.009447111023796929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,1536,0.011758222348160215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,2048,0.3089066611395942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,1536,0.02591733303334978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,1536,0.00868088917599784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,1024,0.008366222182909647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,3072,0.012274666792816587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,1536,0.3857128885057237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,768,0.007724444071451823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,1024,0.00831733312871721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,1024,0.32118222448560924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,768,0.018561777141359117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,768,0.007641777396202087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,512,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,768,0.3284257782830132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,512,0.016217778126398723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,512,0.007289778027269576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,512,0.4046604368421767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,256,0.005666666560702854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,256,0.01574222246805827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,1024,0.020581333173645865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,128,0.004552888787455029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,256,0.36247555414835614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,256,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,128,0.01519377695189582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,64,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,4096,128,0.007268444531493717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,4096,128,0.3239999877081977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,64,0.015185778339703878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,4096,32,0.0148044443792767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,65536,0.26537156105041504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,65536,0.735112879011366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,4096,32,0.004584889031118817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,16384,0.07989955610699125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,65536,0.14153067270914713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,65536,0.9832026163736979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,16384,0.18326044082641602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,12288,0.055585775110456676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,16384,0.357968012491862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,16384,0.03981688949796889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,12288,0.030299554268519085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,12288,0.13546578089396158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,10240,0.04712177647484673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,12288,0.35233155886332196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,8192,0.041464000940322876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,10240,0.11434666315714519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,10240,0.027240888939963445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,10240,0.33151645130581325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,7168,0.03555377655559116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,8192,0.09361244572533502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,8192,0.022656889425383672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,7168,0.08340355422761704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,6144,0.029612445169025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,7168,0.3338177733951145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,6144,0.07234933641221789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,6144,0.32875556415981716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,5120,0.027293332748942908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,6144,0.018596443865034316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,5120,0.06232088804244995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,5120,0.3403502305348714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,5120,0.016685333516862657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,4096,0.0236124445994695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,8192,0.39844799041748047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,7168,0.020747555626763236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,3584,0.021743999587164983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,4096,0.014347554908858405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,4096,0.3349288834465875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,3584,0.012920888761679331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,3072,0.01905155513021681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,3584,0.41069687737358945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,3072,0.03997244437535604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,3072,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,2560,0.017050666941536795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,3072,0.40974312358432347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,4096,0.04933955603175693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,2560,0.03561777869860331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,2048,0.014104000396198697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,2560,0.010655110908879174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,2048,0.029222223493787978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,2560,0.38968533939785427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,3584,0.04458489020665487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,1536,0.011612444288200803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,2048,0.009698666632175446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,2048,0.3135697841644287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,1536,0.02474400069978502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,1536,0.008800889054934183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,1536,0.3814764552646213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,1024,0.01991911066903008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,768,0.007715555528799693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,1024,0.3201795683966743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,768,0.018222222725550335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,768,0.007968000239796108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,768,0.3239857885572645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,512,0.00646666685740153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,512,0.01588444411754608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,1024,0.008743999732865227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,256,0.005232000102599462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,512,0.355879995557997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,512,0.007332444190979004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,256,0.015152888165579902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,256,0.006993778049945831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,128,0.004566222015354368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,1024,0.008176888856622908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,256,0.34214311175876194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,128,0.015104888214005364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,64,0.004267555558019214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3584,32,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3584,128,0.007044444481531779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3584,128,0.3013022210862902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,64,0.0147724449634552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3584,32,0.014815110299322339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,65536,0.19377599822150338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,16384,0.06118488974041409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,65536,0.14479555024041071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,16384,0.17866488297780356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,65536,0.7188142140706381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,16384,0.3577626546223958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,12288,0.047600001096725464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,65536,0.9802133772108289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,12288,0.1333120001686944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,12288,0.030031111505296495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,12288,0.39658665657043457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,10240,0.04087022278043959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,10240,0.11279555161794026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,10240,0.4049822224511041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,16384,0.03825599948565165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,8192,0.03312711252106561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,10240,0.02626488937271966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,8192,0.09223377704620361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,8192,0.021928888228204515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,7168,0.02833333280351427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,8192,0.35966844028896755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,7168,0.0817191137207879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,6144,0.02552177839808994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,7168,0.020299555526839364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,7168,0.35578756862216526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,6144,0.07111555337905884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,5120,0.021202666891945735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,6144,0.018304889400800068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,6144,0.3642595609029134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,5120,0.059544000360700816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,5120,0.016492444607946608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,4096,0.018028444714016385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,5120,0.34517865710788304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,4096,0.04855644371774462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,3584,0.016926222377353244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,4096,0.31276265780131024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,3584,0.04388088981310526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,3072,0.015117333994971381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,3584,0.012847111456924014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,3584,0.31228711869981557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,3072,0.039000888665517174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,2560,0.013231111069520315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,3072,0.41540177663167316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,2560,0.03450577788882785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,2560,0.010693333215183683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,2048,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,4096,0.013592888911565145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,2560,0.3914799955156114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,2048,0.0288702216413286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,1536,0.009268444445398118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,2048,0.32950221167670357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,1536,0.024663110574086506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,3072,0.011776888536082374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,1024,0.007658667034573025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,1536,0.008986666798591614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,1024,0.019542222221692402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,1024,0.34151376618279355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,1024,0.007962666451931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,768,0.006725333217117522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,768,0.01759466694460975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,2048,0.009570666485362584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,768,0.007674666742483775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,512,0.005704888867007361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,512,0.015797333584891427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,1536,0.31986133257548016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,256,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,512,0.007278222176763747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,512,0.40415554576449925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,256,0.014462222655614218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,128,0.004549333204825719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,256,0.007142222589916653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,256,0.3269386556413439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,128,0.014446222119861178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,64,0.004560889055331548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,3072,128,0.0069448889957533936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,3072,32,0.004871110949251386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,768,0.3329857720269097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,3072,128,0.3692604435814752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,64,0.014146667387750415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,3072,32,0.014154665999942355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,65536,0.1835404502020942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,16384,0.04942755566702949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,65536,0.1427031093173557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,16384,0.16914667023552787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,65536,0.6762488683064779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,12288,0.03936800029542711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,16384,0.037552889850404524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,65536,0.9803955290052625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,12288,0.329947551091512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,10240,0.034075554874208235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,12288,0.02835822105407715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,10240,0.10616799857881333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,10240,0.331803560256958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,10240,0.025399999486075506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,8192,0.02719466719362471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,16384,0.331458674536811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,8192,0.08746400144365098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,7168,0.024442666106753882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,8192,0.020972443951500788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,12288,0.12539645036061606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,7168,0.07704444726308186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,6144,0.021228444245126512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,7168,0.01921688848071628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,7168,0.32958756552802193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,6144,0.06685155630111694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,6144,0.01733955575360192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,5120,0.01940088967482249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,6144,0.45981778038872617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,5120,0.05597866906060112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,8192,0.32454223102993435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,4096,0.015644444359673392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,5120,0.01590577761332194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,5120,0.4351786772410075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,4096,0.04660177893108792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,3584,0.014217777384652032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,4096,0.013372444444232516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,4096,0.36008887820773655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,3584,0.041165331999460854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,3072,0.012548444171746572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,3584,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,3584,0.38580799102783203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,3072,0.037032001548343234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,2560,0.011191999746693505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,3072,0.011428444749779172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,3072,0.3551528983645969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,2560,0.032891554964913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,2048,0.00944977750380834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,2560,0.010599111517270407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,2560,0.4009564452701145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,2048,0.362847990459866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,2048,0.009376889301670922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,1536,0.008338666624493068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,1536,0.023334221707450017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,1536,0.3544746769799127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,1024,0.006879111131032308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,1536,0.00870400004916721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,1024,0.018908444378111098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,2048,0.027515555421511333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,768,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,1024,0.41637243164910215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,768,0.01711466742886437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,768,0.007681777907742395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,512,0.00535733335547977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,768,0.3312275674608019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,512,0.015491555134455362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,512,0.007278222176763747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,256,0.004609777695602841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,256,0.014130666851997375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,512,0.42653510305616593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,128,0.004266666869322459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,256,0.0069475554757648045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,256,0.38993332121107316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,1024,0.007954667011896769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,64,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2560,128,0.00703022215101454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2560,32,0.003946666502290302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2560,128,0.30602576997545033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,64,0.013801777528391944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,65536,0.15564089351230198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,32,0.013824889229403602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,65536,0.09359288877911037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,16384,0.047714667187796704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,65536,0.6328586472405328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2560,128,0.014130666851997375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,16384,0.15960355599721274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,16384,0.3102960056728787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,12288,0.03449777762095133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,65536,0.8449084493849012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,16384,0.026258667310078938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,10240,0.02959733208020528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,12288,0.11839289135403103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,10240,0.09976977772182888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,12288,0.39002309905158145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,8192,0.024286222126748826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,10240,0.32426222165425617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,10240,0.01866222255759769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,8192,0.0810497800509135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,8192,0.015660444895426433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,7168,0.021406221720907424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,8192,0.3237280050913493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,7168,0.07243733273612128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,7168,0.014855111638704935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,6144,0.018460444278187223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,7168,0.37891289922926163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,12288,0.020404444800482858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,6144,0.06266488631566365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,5120,0.01649599936273363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,6144,0.013550221920013428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,6144,0.3259546756744385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,4096,0.014179555906189812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,5120,0.05212355653444926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,5120,0.012293332980738746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,5120,0.32923732863532174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,3584,0.012857777376969656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,4096,0.04304800099796719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,4096,0.010724444356229572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,4096,0.3180008994208442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,3072,0.011658666862381829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,3584,0.03936533464325799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,3584,0.010017777482668558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,3584,0.3288782172732883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,2560,0.01016088906261656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,3072,0.035353776481416486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,3072,0.009444444543785518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,3072,0.31036443180508083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,2560,0.0302817788388994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,2048,0.00814133303032981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,2560,0.008656000097592672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,2560,0.3474862310621474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,2048,0.02611911131276025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,1536,0.007121777368916406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,2048,0.3538382318284776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,1536,0.022277333670192297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,1536,0.007638221813572778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,1024,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,1536,0.3817431131998698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,1024,0.01826844447188907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,1024,0.35944800906711155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,1024,0.0070497774415545994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,768,0.005247999810510212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,2048,0.008310221963458592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,768,0.016549333930015564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,512,0.004967999955018361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,768,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,768,0.34688533676995176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,512,0.014911999305089315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,256,0.004252444538805219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,512,0.006616000086069107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,512,0.34155111842685276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,256,0.014186667071448432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,128,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,256,0.006432000133726332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,256,0.3530720074971517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,128,0.013519110778967539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,64,0.0038720000949170855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,2048,128,0.006254222244024277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,2048,32,0.0038764443662431505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,2048,128,0.3177439901563856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,64,0.013467555244763693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,2048,32,0.013078221844302284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,65536,0.1334177785449558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,16384,0.03866666555404663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,65536,0.09062755770153469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,65536,0.5926106770833334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,16384,0.3257066673702664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,16384,0.14880177709791395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,12288,0.028568890359666612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,16384,0.025221332907676697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,65536,0.8323760032653809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,12288,0.020392888122134738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,10240,0.02346399923165639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,12288,0.10977333121829563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,12288,0.325077321794298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,8192,0.020267556111017864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,10240,0.09280444516075982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,10240,0.3116950988769531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,10240,0.01811911165714264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,7168,0.017839999662505258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,8192,0.076837334367964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,8192,0.015146666102939181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,8192,0.3225395679473877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,6144,0.015875554747051664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,7168,0.013769778112570444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,7168,0.3286657863193088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,6144,0.0559084415435791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,5120,0.013755555782053204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,6144,0.013069333301650153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,6144,0.3266515466901991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,5120,0.0492248899406857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,4096,0.011910222470760345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,5120,0.012048888537618848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,5120,0.3274631235334608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,4096,0.04049955474005805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,3584,0.01015644437736935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,4096,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,4096,0.3127315574222141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,3584,0.036687112516827054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,3584,0.009650666680600908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,7168,0.06759555472267999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,3584,0.32832177480061847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,3072,0.03232533401913113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,2560,0.008643555144468943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,3072,0.009359111388524374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,3072,0.35831112331814235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,2560,0.02922755479812622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,2048,0.00741244438621733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,2560,0.31625244352552623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,2560,0.0087333329849773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,2048,0.024953777591387432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,1536,0.006604444649484422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,2048,0.008289777570300633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,2048,0.30915021896362305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,1536,0.020996444755130347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,3072,0.00944266633854972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,1024,0.005408888889683618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,1536,0.007646222081449296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,1024,0.01752444439464145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,768,0.004938666605287128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,1024,0.32676622602674693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,768,0.015876443849669564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,768,0.006640000061856375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,768,0.32507022221883136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,512,0.004592888885074192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,512,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,512,0.42730755276150173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,1536,0.3846435546875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,512,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,256,0.0041759999261962045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,256,0.01348000019788742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,256,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,256,0.4277786678738064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,128,0.0041911109454101985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,128,0.013188444077968597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1536,128,0.3384888966878255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,64,0.003919110943873723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,128,0.006334222025341458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1536,32,0.0039048890272776284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1536,1024,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,64,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,65536,0.09111733569039239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1536,32,0.012829333543777466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,65536,0.08835466702779134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,16384,0.03284800052642822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,65536,0.5501946873135036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,65536,0.8237502310011121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,16384,0.13917156060536703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,12288,0.023641778363121882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,16384,0.36670488781399196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,16384,0.024777778320842322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,12288,0.10258577929602729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,12288,0.019208888212839764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,10240,0.0207315550910102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,10240,0.08657511075337727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,10240,0.01752355529202355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,8192,0.017482666505707633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,8192,0.07117066780726115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,8192,0.014808000789748298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,7168,0.01590844492117564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,8192,0.479390197330051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,12288,0.3577724562750922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,7168,0.06295022037294176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,7168,0.013763555222087435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,6144,0.013873777455753751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,10240,0.38878220982021755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,7168,0.40082399050394696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,6144,0.053048001395331494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,5120,0.01180266671710544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,6144,0.3238862090640598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,6144,0.012810666528013019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,5120,0.04473955432573954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,4096,0.010430222584141625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,5120,0.012098666694429187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,5120,0.3577271037631565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,3584,0.009189333352777693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,4096,0.01036622209681405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,4096,0.40862311257256406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,3584,0.03350399931271871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,3072,0.00830577810605367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,3584,0.325962675942315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,3584,0.00980177770058314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,3072,0.029523554775449965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,2560,0.007634667058785756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,3072,0.009416888985368941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,3072,0.32293065388997394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,2560,0.02681155502796173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,2048,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,2560,0.00868444475862715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,4096,0.037407110134760536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,2560,0.36572265625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,2048,0.023166222704781428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,1536,0.005931555396980709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,2048,0.007989333735571967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,2048,0.3215840127733019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,1024,0.005229333208666907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,1536,0.007605333295133378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,1536,0.3621946705712213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,1024,0.016557332542207506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,768,0.004886222382386525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,1024,0.00702133360836241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,768,0.015491555134455362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,768,0.0068195557428730865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,512,0.004210666649871402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,768,0.39382044474283856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,512,0.013888888888888888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,1536,0.01997066703107622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,512,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,256,0.003924444317817688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,512,0.38367732365926105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,256,0.01314933349688848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,128,0.003855111284388436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,256,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,1024,0.31263377931382924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,256,0.33754399087693954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,64,0.0038533334930737815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,128,0.012815111213260226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,1024,128,0.006285333385070165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,64,0.012413333687517377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,1024,32,0.012488889197508493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,65536,0.0772835546069675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,65536,0.5442710982428657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,65536,0.08747022019492255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,1024,32,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,16384,0.026484444737434387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,1024,128,0.36548354890611434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,16384,0.137807117568122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,16384,0.35625243186950684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,12288,0.018577777677112155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,16384,0.024131556351979572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,12288,0.10160355435477363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,12288,0.3694053226047092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,12288,0.019260444574885897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,10240,0.016344888342751395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,65536,0.8212622006734213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,10240,0.08558311065038045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,10240,0.016906667086813185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,8192,0.013557333085272046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,10240,0.35198399755689835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,8192,0.06950577762391832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,8192,0.014847111370828418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,7168,0.012302222351233164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,8192,0.33164800537957084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,7168,0.06097866429222954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,7168,0.013742222554153867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,6144,0.011349333657158745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,7168,0.35124890009562176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,6144,0.05212000012397766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,5120,0.00981866651111179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,6144,0.012814222110642327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,6144,0.3577413294050429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,5120,0.04409689042303297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,5120,0.011863110793961419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,5120,0.3661048942142063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,4096,0.037348445918824934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,4096,0.010422222316265106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,4096,0.3552551004621718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,3584,0.03400977783732944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,3584,0.009681777821646797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,3072,0.006996444529957241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,4096,0.008375999828179678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,3072,0.029771556456883747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,3584,0.553244431813558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,3072,0.3162062168121338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,3072,0.008995555341243744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,2560,0.006711999989218182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,3584,0.007903111477692923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,2048,0.005985777825117111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,2560,0.026857778429985046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,2560,0.008725333544943068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,2560,0.3241795433892144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,1536,0.005249777601824866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,2048,0.023339556323157415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,2048,0.3210008939107259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,2048,0.008062221937709385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,1536,0.01958933307064904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,1024,0.004625777817434735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,1536,0.00739555557568868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,1536,0.32430577278137207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,1024,0.016164445214801364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,1024,0.007027555671003129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,768,0.014754666222466363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,1024,0.452272891998291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,512,0.004287999951177173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,768,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,768,0.31719912423027885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,512,0.006577777779764599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,512,0.013850666582584381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,256,0.003572444534964032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,512,0.32293865415785045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,256,0.013087111214796701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,256,0.006254222244024277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,128,0.003644444462325838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,256,0.3259493245018853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,768,0.004267555558019214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,64,0.0035413333939181436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,768,128,0.006370666540331311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,768,32,0.00360000009338061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,64,0.01240533341964086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,768,128,0.4012320041656494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,32,0.0123831108212471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,65536,0.5034489101833767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,768,128,0.012420444852775998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,65536,0.08789155880610149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,16384,0.019334221879641216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,65536,0.05411555700831943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,16384,0.12831021679772273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,16384,0.3281244436899821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,12288,0.017694221602545846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,16384,0.023780445257822674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,12288,0.09303555885950725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,12288,0.018960000740157235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,10240,0.015509333875444202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,12288,0.4090400007035997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,10240,0.07911466889911227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,8192,0.013150222599506378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,65536,0.82132355372111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,10240,0.01696888936890496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,8192,0.06409421894285414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,7168,0.011855111353927188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,8192,0.014621333943472968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,8192,0.32185599539015025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,7168,0.05579377545250786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,6144,0.010115555591053432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,7168,0.39911288685268825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,6144,0.04829066660669115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,10240,0.32458578215705025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,6144,0.458679993947347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,5120,0.009232000344329411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,6144,0.012466666599114736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,5120,0.040836443503697716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,7168,0.01371911085314221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,4096,0.008050666915045844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,5120,0.011784888803958893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,4096,0.03463110989994473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,4096,0.01034311122364468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,3584,0.007584888901975419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,4096,0.4673742188347711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,3584,0.030879110097885132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,3584,0.009712888962692684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,3072,0.00664444433318244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,5120,0.37815377447340226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,3072,0.02786133355564541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,3072,0.3454391161600749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,2560,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,2560,0.025421332981851365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,3584,0.4353200064765082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,2560,0.4019333256615533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,2048,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,2560,0.00867022242810991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,2048,0.021691555778185528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,3072,0.00905600024594201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,1536,0.004884444591071871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,1536,0.018939556346999276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,2048,0.47082842720879453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,1536,0.32474666171603733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,1024,0.004368000146415499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,1536,0.007428444094128079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,1024,0.015565334094895257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,1024,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,768,0.003944000022278892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,768,0.014527999692493014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,1024,0.4167582194010417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,512,0.003871110992299186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,768,0.31100355254279244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,2048,0.007989333735571967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,512,0.013183111117945777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,256,0.003581333491537306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,512,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,512,0.31940799289279515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,256,0.012467555701732635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,128,0.003261333331465721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,256,0.3853902286953396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,768,0.00665511108107037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,512,128,0.357638226615058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,64,0.0032311110860771606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,128,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,512,32,0.0031617778456873367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,64,0.012437333663304647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,32,0.012081777883900536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,65536,0.03648533423741659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,512,256,0.006275555739800136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,512,128,0.012103999654452005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,65536,0.08717511097590129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,16384,0.013583111266295115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,65536,0.5026826858520508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,16384,0.12703022691938612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,12288,0.011046222514576383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,16384,0.022988445229000513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,16384,0.3497680028279622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,65536,0.8201653162638346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,10240,0.011838222543398539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,12288,0.018808888064490426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,12288,0.3240577909681532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,10240,0.07891111241446601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,8192,0.012223111258612739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,10240,0.016880000631014507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,10240,0.3475484318203396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,8192,0.06330933173497517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,8192,0.01459022197458479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,7168,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,8192,0.322160005569458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,7168,0.05580266979005602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,7168,0.013521778086821238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,6144,0.010113777385817634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,7168,0.39182400703430176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,6144,0.04850577645831638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,12288,0.09183200200398763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,5120,0.008992888861232335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,6144,0.3330026732550727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,6144,0.012647111382749347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,5120,0.03940088881386651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,4096,0.00796977761718962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,5120,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,5120,0.3259066740671794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,3584,0.007275555696752336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,4096,0.010374222364690568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,3584,0.03103733393881056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,3584,0.009696000152164036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,3072,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,3072,0.027858667903476294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,3072,0.3443022304111057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,4096,0.033137778441111244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,2560,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,3072,0.009102221992280748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,4096,0.37933866182963055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,2560,0.02477955487039354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,3584,0.33377689785427517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,2560,0.00833777752187517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,2048,0.00555644432703654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,2560,0.3398711151546902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,2048,0.021378666162490845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,1536,0.00489155575633049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,2048,0.3472862243652344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,1536,0.018546667363908555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,1536,0.007420444654093848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,1536,0.3536649015214708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,1024,0.0041404445138242514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,1024,0.015541333291265698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,1024,0.3617253303527832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,1024,0.006959111326270633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,768,0.0038862224254343244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,2048,0.00793333351612091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,768,0.014467555615637036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,768,0.006632888896597757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,512,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,768,0.33817778693305117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,512,0.013066666821638743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,256,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,512,0.006314666734801398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,512,0.33922844462924534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,256,0.012413333687517377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,128,0.00309244439833694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,256,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,256,0.338847107357449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,128,0.01202933324707879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,64,0.0029146667155954572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,256,128,0.00630666646692488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,256,128,0.31385776731703013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,256,32,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,64,0.011742221812407175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,65536,0.02846577763557434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,256,32,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,16384,0.010431110858917236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,65536,0.0878231128056844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,65536,0.5026115311516656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,16384,0.023012444376945496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,16384,0.3266071213616265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,16384,0.12600978215535483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,65536,0.8200142118665906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,12288,0.009032888544930352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,10240,0.011117333339320289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,12288,0.018616888258192275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,12288,0.09264177746242946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,10240,0.07885511053933038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,8192,0.010018666585286459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,10240,0.016955556141005624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,10240,0.3131768968370226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,12288,0.5286622047424316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,8192,0.06382489204406738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,7168,0.009512888888518015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,8192,0.014436443646748861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,8192,0.3262844350602892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,6144,0.008669333325492011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,7168,0.05458577805095249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,7168,0.013431111143694984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,6144,0.04670844475428263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,6144,0.012392889294359418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,5120,0.008043555749787224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,7168,0.44690577189127606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,6144,0.30735911263359916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,4096,0.007649777664078607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,5120,0.03922222057978312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,5120,0.011791999969217511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,4096,0.0333262218369378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,3584,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,4096,0.010437332921557956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,4096,0.3254515594906277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,3584,0.03055022160212199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,3072,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,3584,0.009361777868535783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,3072,0.0277146663930681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,3584,0.4075031015608046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,2560,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,3072,0.3201679918501112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,2560,0.024944000773959692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,2560,0.3145955668555366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,5120,0.31087199846903485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,2048,0.005548444473081165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,2560,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,2048,0.02161600026819441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,1536,0.004794666750563515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,2048,0.31390221913655597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,1536,0.018931556079122756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,1536,0.007460444337791867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,1536,0.31956177287631565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,1024,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,1024,0.01514311134815216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,1024,0.3315075503455268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,1024,0.006950221955776215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,768,0.0038693332009845306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,768,0.014177777700954013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,2048,0.008046222229798635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,512,0.0035582222044467926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,768,0.31874312294854057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,768,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,512,0.013152889079517789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,3072,0.009008000294367472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,512,0.34224266476101345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,256,0.0035031110876136353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,512,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,256,0.012123555772834353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,256,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,128,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,256,0.3613751199510362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,128,0.011818666425016193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,64,0.0028719999310043124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1024,128,128,0.006239111224810283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,1024,128,128,0.33829156557718915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,128,32,0.003175111073586676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,65536,0.02720622221628825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,64,0.01181333346499337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,128,32,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,16384,0.00960533320903778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,12288,0.008396444221337637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,16384,0.1252231068081326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,65536,0.5038871235317653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,12288,0.09160266982184516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,8192,0.008307555483447181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,10240,0.07830133040746053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,7168,0.007770666645632849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,8192,0.06308711237377591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,6144,0.007393777370452881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,7168,0.0541928874121772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,5120,0.007639110916190677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,6144,0.04652977652019925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,4096,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,5120,0.03963199920124478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,3584,0.006973333656787872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,10240,0.008810666700204214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,3072,0.006340444501903322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,4096,0.03315733207596673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,3584,0.030518223841985066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,2560,0.005679111099905438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,2048,0.005280888742870755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,3072,0.027492443720499676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,2560,0.025063110722435847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,1536,0.004606222112973531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,1024,0.003903111235962974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,2048,0.021706667211320665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,768,0.003907555507289039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,1536,0.018575999471876357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,512,0.0035093331502543557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,1024,0.01550222271018558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,256,0.003272888975010978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,768,0.014514666464593677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,512,0.012787555654843649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,128,0.0029324444217814338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,64,0.0029191111938820947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,64,32,0.0028248888750871024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,128,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,65536,0.025495110286606684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,32,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,64,0.011543110840850405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,16384,0.011698666546079846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,12288,0.01037688884470198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,16384,0.1259431176715427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,12288,0.0933377808994717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,10240,0.009370666411187913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,8192,0.008751110898123847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,65536,0.5018666585286459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,7168,0.008381333616044786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,10240,0.07829244269265069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,6144,0.008046222229798635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,8192,0.06242044766743978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,5120,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,7168,0.054140445258882314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,64,256,0.012105777859687805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,6144,0.04718666606479221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,4096,0.007339555356237624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,3584,0.0069413334131240845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,3072,0.006611555400821898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,5120,0.03891377647717794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,4096,0.03316355413860745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,2560,0.005963555640644497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,3584,0.03049066662788391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,3072,0.027525333894623652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,1536,0.005230222311284807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,2560,0.02500444485081567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,1024,0.0041955556306574075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,2048,0.021702221698231165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,1536,0.018900444110234577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,768,0.0038897775941424896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,512,0.0035911111368073356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,1024,0.015461333923869662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,256,0.0032115555885765287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,768,0.014147554834683737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,128,0.0031831111345026228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,512,0.012829333543777466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,256,0.01237155579858356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,128,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,32,0.0028275555620590844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,64,0.011773332953453064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1024,32,32,0.011695110963450538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,2048,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,16384,0.821604463789198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1024,32,64,0.0029084444459941653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,16384,1.0825733608669705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,16384,0.2834764321645101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,12288,0.7394008636474609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,12288,0.7752284473843045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,16384,1.1085039774576824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,12288,0.21696800655788848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,10240,0.7003839810689291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,12288,0.7792186737060547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,10240,0.6437884436713325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,10240,0.18750577502780488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,8192,0.44008710649278426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,10240,0.609640015496148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,8192,0.16197689374287924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,8192,0.5370924207899306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,7168,0.5272409121195475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,7168,0.44557513131035703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,8192,0.5185004340277778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,7168,0.13211111227671304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,6144,0.33605334493849015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,7168,0.4633750915527344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,6144,0.392145766152276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,5120,0.29770488209194607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,6144,0.11881511741214329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,5120,0.32567556699117023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,5120,0.10255289077758789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,4096,0.22807910707261828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,5120,0.3524613380432129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,4096,0.2636275556352404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,3584,0.19360355536142984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,4096,0.39039821094936794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,4096,0.08631910880406697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,3584,0.22961333062913683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,3584,0.07776000102361043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,6144,0.4013813336690267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,3072,0.16649688614739314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,3584,0.3675599892934163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,3072,0.199346661567688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,2560,0.15188444985283747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,3072,0.30797600746154785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,2560,0.1703039937549167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,2048,0.11710400051540798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,2560,0.061575108104281955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,2560,0.34668445587158203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,2048,0.14129422770606145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,1536,0.10157422224680583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,2048,0.3246844344668918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,1536,0.1111724509133233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,3072,0.06959733035829332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,1536,0.31785066922505695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,1024,0.07327377796173096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,1536,0.04294311006863912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,1024,0.08584355645709568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,1024,0.035601778162850276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,768,0.059373332394493945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,1024,0.3891119956970215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,768,0.07122044430838691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,512,0.04063377777735392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,2048,0.054224888483683266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,768,0.3217804431915283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,768,0.03215377860599094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,256,0.033412443266974554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,512,0.029121776421864826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,512,0.30402933226691353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,256,0.05705066521962484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,128,0.023375110493765935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,256,0.027106665902667578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,256,0.32435909907023114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,128,0.053744888967937894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,64,0.020388444264729817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,65536,128,0.29180977079603404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,65536,32,0.021057777934604224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,64,0.05288177728652954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,512,0.06211022535959879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,65536,32,0.05316799879074097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,65536,0.8791031307644315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,65536,128,0.0266977780395084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,16384,0.22583643595377603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,65536,1.3014613257514107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,16384,0.3176497883266873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,65536,0.3501359886593289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,12288,0.18285689089033338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,16384,0.3853448761834039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,12288,0.23811822467380098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,12288,0.3095217810736762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,12288,0.06857777966393365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,65536,1.3844772974650066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,10240,0.16016177336374918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,16384,0.08541155523724026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,10240,0.058479110399881996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,8192,0.12213599681854248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,10240,0.329984876844618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,8192,0.16702400313483345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,7168,0.10823466380437215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,8192,0.04978488882382711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,8192,0.3362729019588894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,7168,0.14356444941626653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,7168,0.042913777960671314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,6144,0.09412799941168891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,7168,0.3268311023712158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,10240,0.20371733771430123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,6144,0.12515910466512045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,5120,0.07594755623075697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,6144,0.03977333174811469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,6144,0.35138577885097927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,5120,0.10804444551467896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,5120,0.03446577654944526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,4096,0.061795552571614586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,5120,0.41294132338629824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,4096,0.08843644460042317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,3584,0.056072890758514404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,4096,0.37308088938395184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,4096,0.02792888879776001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,3584,0.08069333102968004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,3584,0.024930665890375774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,3072,0.04862577716509501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,3584,0.36034843656751847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,3072,0.07016355461544461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,3072,0.022241777843899194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,3072,0.3703911039564345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,2560,0.06028088596132067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,2560,0.36577865812513566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,2048,0.03407288922203912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,2048,0.050945778687795006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,2560,0.040580444865756564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,2048,0.377670235104031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,1536,0.024710221423043147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,2048,0.01756622228357527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,2560,0.020023110840055678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,1536,0.04047466648949517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,1024,0.018431999617152743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,1536,0.015000000596046448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,1024,0.031910220781962075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,1536,0.3990800115797255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,768,0.014353776971499125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,1024,0.012599111431174807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,1024,0.3238897853427463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,512,0.013195555243227216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,768,0.02759733299414317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,768,0.01146933353609509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,768,0.3223288853963216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,512,0.02368800010946062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,256,0.010763555765151978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,512,0.010785777535703448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,256,0.02163911031352149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,512,0.3860924508836534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,256,0.31396354569329155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,256,0.010428444378905825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,128,0.021010667085647583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,64,0.009514666265911525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,16384,128,0.010438222024175854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,32,0.009697777529557547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,64,0.020949333906173706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,16384,32,0.020671111014154222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,16384,128,0.01000711073478063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,65536,0.5655680232577854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,16384,128,0.2941208945380317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,65536,1.051511976453993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,16384,0.1554995510313246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,65536,0.2473511166042752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,16384,0.2596497800615099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,16384,0.06687466965781318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,12288,0.1154151095284356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,16384,0.3251288996802436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,12288,0.19689599672953287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,10240,0.10537511110305786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,12288,0.3117608759138319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,10240,0.16377333799997965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,65536,1.0993262396918404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,10240,0.3057759867774116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,8192,0.09065155850516425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,10240,0.045349333021375865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,8192,0.1373128890991211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,7168,0.07758044534259372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,8192,0.03741155399216546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,8192,0.3315253257751465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,7168,0.11850399441189235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,7168,0.033182223637898765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,6144,0.07057422399520874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,7168,0.3989866574605306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,6144,0.10441866848203872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,5120,0.05961866511238945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,6144,0.029851555824279785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,6144,0.334154658847385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,5120,0.09043466382556492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,5120,0.025835555460717943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,4096,0.0481333335240682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,12288,0.05177422364552816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,5120,0.36377867062886554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,3584,0.04277955492337545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,4096,0.0754293335808648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,4096,0.021377777059872944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,4096,0.31258400281270343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,3072,0.036770668294694685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,3584,0.06732177734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,3584,0.3272133403354221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,3072,0.05619733201132881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,2560,0.03143644332885742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,3072,0.01770400007565816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,3072,0.3901813295152452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,2560,0.05051911208364698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,2560,0.3556515640682644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,2560,0.015819556183285184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,2048,0.026867555247412786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,2048,0.04320977793799507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,3584,0.01944266590807173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,2048,0.3679537773132324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,2048,0.014284445179833306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,1536,0.034486220942603216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,1536,0.3823386563195123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,1536,0.012991111311647626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,1024,0.015480000111791821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,1024,0.028018666638268366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,1024,0.3663253254360623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,1024,0.010879110958841113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,768,0.01295111080010732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,1536,0.021711111068725586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,768,0.024449777272012498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,512,0.010000000397364298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,768,0.009717333647939894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,768,0.34608088599310977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,512,0.020710221595234342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,256,0.0069431112044387394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,512,0.009310222334331935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,512,0.33243465423583984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,256,0.019296889503796894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,128,0.006583111153708563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,256,0.008712000317043727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,256,0.3370560010274251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,128,0.018930666976504855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,12288,128,0.008939555949634975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,64,0.005931555396980709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,12288,32,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,12288,128,0.3144959873623318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,64,0.018564444449212816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,12288,32,0.018538667096032035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,65536,0.5257617632548014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,16384,0.16051999727884927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,65536,0.24292800161573622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,65536,0.9408017264472114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,16384,0.2309724489847819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,12288,0.11265422238243951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,16384,0.06218577755822075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,12288,0.17545689476860893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,16384,0.5265937911139594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,12288,0.3261840078565809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,10240,0.08974933624267578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,12288,0.0490515563223097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,65536,1.1021075778537326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,10240,0.04325422313478258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,10240,0.1485395563973321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,8192,0.07703377803166707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,8192,0.122707552380032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,10240,0.46920177671644425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,8192,0.03602222270435757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,7168,0.06421599785486858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,8192,0.3182906574673123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,6144,0.05503466725349426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,7168,0.032003555032942034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,7168,0.10714577966266209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,7168,0.3244222270117866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,6144,0.09348711040284897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,5120,0.04722399844063652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,6144,0.02808444367514716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,6144,0.3256559901767307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,5120,0.07907822397020128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,4096,0.03925688730345832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,5120,0.024223110742039148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,5120,0.3288968933953179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,3584,0.03453333179155985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,4096,0.06647288799285889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,4096,0.020671111014154222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,4096,0.32320621278550893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,3072,0.030241777499516804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,3584,0.059635553095075816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,3584,0.018767111831241183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,3072,0.05198311143451267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,2560,0.026572444372706946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,3072,0.3265484439002143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,3072,0.0173688895172543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,2560,0.043976889716254346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,2048,0.021672889590263367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,2560,0.015807999504937064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,2560,0.3185066646999783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,2048,0.038607110579808555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,1536,0.01793333391348521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,2048,0.01388177772363027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,2048,0.4052551057603624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,1536,0.03107200066248576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,1024,0.012845333251688214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,1536,0.28940267033047146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,3584,0.3282097710503472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,1536,0.012283555335468717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,1024,0.02541422181659275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,768,0.011438222395049201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,1024,0.01016266644001007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,768,0.02167911165290409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,1024,0.30497601297166615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,512,0.00906755526860555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,768,0.009619555539555019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,512,0.019347555107540555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,512,0.3533626662360297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,512,0.008990222381220924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,256,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,256,0.01819022165404426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,256,0.3645377688937717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,256,0.008648888932334052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,128,0.0063324446479479475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,128,0.017701332767804463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,128,0.3310640123155382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,10240,128,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,64,0.007688889073001013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,10240,32,0.00794488853878445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,64,0.017535999417304993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,10240,32,0.01719199948840671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,65536,0.42537689208984375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,10240,768,0.4041786723666721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,65536,0.8282657729254829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,16384,0.11328978008694118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,65536,0.19140177302890352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,16384,0.203222221798367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,16384,0.05217155483033922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,12288,0.09086400270462036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,16384,0.41776267687479657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,65536,0.9599564870198568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,12288,0.15554932753245035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,10240,0.0673253337542216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,12288,0.0413955549399058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,12288,0.3837760024600559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,10240,0.13404177294837102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,8192,0.06242577897177803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,10240,0.03544977638456557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,8192,0.10917422506544326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,10240,0.45482757356431747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,7168,0.04907822277810839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,8192,0.38034489419725204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,7168,0.0958942241138882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,6144,0.04240266813172234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,7168,0.3360648949940999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,7168,0.025497777594460383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,6144,0.02307022280163235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,6144,0.08385955625110203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,5120,0.04087911049524943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,6144,0.3322790993584527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,5120,0.07003821929295857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,5120,0.020527111159430612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,8192,0.02886311213175456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,4096,0.029714667134814795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,5120,0.3990444342295329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,4096,0.05859466393788656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,3584,0.02935289012061225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,4096,0.32622843318515354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,3584,0.05255377623769972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,3072,0.026233777403831482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,3584,0.01629866659641266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,3584,0.39070844650268555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,3072,0.046147555112838745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,2560,0.022335110439194575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,3072,0.014868444866604276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,3072,0.32815466986762154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,2560,0.04084088736110263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,2048,0.019280888968043856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,2560,0.013527111046844058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,2048,0.03513866662979126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,2560,0.37861156463623047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,4096,0.01756533318095737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,1536,0.015544000599119397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,2048,0.32799821429782444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,2048,0.011790222591824003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,1024,0.011756444142924415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,1536,0.028539554940329656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,1536,0.010482666393121084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,1024,0.02292088833120134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,768,0.009340444372759925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,1024,0.00924711094962226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,1024,0.3535439968109131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,768,0.02000799940692054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,512,0.007661333514584436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,768,0.008638222184446123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,768,0.32535110579596627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,512,0.01791200041770935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,256,0.0069164443347189165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,512,0.00832711077398724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,512,0.33021865950690377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,256,0.016865778300497267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,128,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,256,0.007948444121413762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,256,0.32711733712090385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,128,0.016883555385801528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,1536,0.3758186764187283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,64,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,8192,32,0.006582222051090664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,8192,128,0.008003555238246918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,8192,128,0.30476533042060006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,64,0.01622222198380364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,8192,32,0.016142222616407607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,65536,0.33742488755120176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,16384,0.09585689173804389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,65536,0.18999377886454263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,16384,0.20824800597296822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,65536,0.7656168937683105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,16384,0.3431235684288873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,16384,0.05013066530227661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,12288,0.07237955596711901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,12288,0.1420079999499851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,65536,0.9579209221733941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,12288,0.3691253397199843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,10240,0.06643378072314791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,12288,0.03862844573126899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,10240,0.12149066395229763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,10240,0.033350222640567355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,8192,0.05334933267699348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,10240,0.4163520071241591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,8192,0.0984613365597195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,7168,0.04212711254755656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,8192,0.36117511325412327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,8192,0.02792888879776001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,7168,0.08755111032062107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,7168,0.025036444266637165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,6144,0.0369120008415646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,7168,0.37014044655693906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,6144,0.07740355862511529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,6144,0.023028444912698533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,5120,0.03492800063557095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,6144,0.356312010023329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,5120,0.06474755869971381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,4096,0.029412445094850328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,5120,0.020006222857369315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,5120,0.4162453280554877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,4096,0.053335110346476235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,3584,0.026019554999139573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,4096,0.34581687715318465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,4096,0.017367111312018502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,3584,0.04799911048677233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,3072,0.022572444544898138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,3584,0.016025778320100572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,3584,0.3283253245883518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,3072,0.042071110672420926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,2560,0.019810666640599568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,3072,0.014117333624098035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,3072,0.35115554597642684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,2048,0.016898666818936665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,2560,0.03729155659675598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,2048,0.03279022375742594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,2048,0.31187110477023655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,2560,0.48736980226304794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,1536,0.013766222529941134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,2048,0.011759999725553723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,1536,0.026963555150561865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,1024,0.009935110807418823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,1536,0.31895645459493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,1024,0.0214035560687383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,1024,0.3172053231133355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,2560,0.012964444855848948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,768,0.008960000342792934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,1024,0.009023110899660323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,768,0.01920622256067064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,512,0.006959999601046245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,768,0.00867377801073922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,768,0.3188542260064019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,512,0.017190222938855488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,256,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,512,0.008032889001899296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,512,0.368485344780816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,256,0.016152888536453247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,128,0.0052622221410274506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,256,0.008002666963471307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,128,0.015825778245925903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,128,0.007678222325113084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,128,0.33937867482503253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,64,0.004588444613748127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,7168,32,0.004538666870858935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,64,0.0158595558669832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,7168,32,0.015484443969196744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,7168,1536,0.010031110710567897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,65536,0.3068319956461589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,7168,256,0.32455378108554417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,16384,0.08547555738025242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,65536,0.14732889334360758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,65536,0.7068826887342664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,16384,0.17512710889180502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,12288,0.06703466839260526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,16384,0.03952266772588094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,16384,0.3366542127397325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,12288,0.13389510578579372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,12288,0.031416889694001936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,10240,0.05762133333418104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,12288,0.3244933287302653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,65536,0.8249955707126193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,8192,0.04760266674889458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,10240,0.026318222284317017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,10240,0.11402666568756104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,10240,0.32295555538601345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,7168,0.04064088728692796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,8192,0.022654222117529974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,8192,0.0932275586658054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,8192,0.31733155250549316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,6144,0.03471555643611484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,7168,0.08285422457589044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,7168,0.02014844450685713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,7168,0.34056711196899414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,6144,0.0714942216873169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,5120,0.028982222080230713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,6144,0.018432888719770644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,5120,0.059917330741882324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,5120,0.37298043568929035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,5120,0.01640888883007897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,4096,0.02563200063175625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,4096,0.049045334259668984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,4096,0.37636354234483504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,3584,0.023223999473783705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,6144,0.35114044613308376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,3584,0.04504088891877068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,3584,0.013208889298968844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,3584,0.37213601006401914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,3072,0.020872889293564692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,3072,0.039846221605936684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,3072,0.3620755672454834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,3072,0.011770666473441653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,2560,0.01793955597612593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,4096,0.014230221509933472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,2560,0.035487999518712364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,2048,0.015097777048746744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,2560,0.010468444062603844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,2560,0.3534879949357774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,2048,0.030751112434599135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,1536,0.012699555191728802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,2048,0.009959999885823991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,2048,0.3497457769181993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,1536,0.025014221668243408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,1024,0.009843555589516958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,1536,0.008736888567606607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,1536,0.3496239980061849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,768,0.00813066628244188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,1024,0.020279111133681405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,1024,0.32226488325330943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,1024,0.008282666405042013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,512,0.006903111106819577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,768,0.329382234149509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,768,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,512,0.007292444507280986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,256,0.005616000129116907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,512,0.32353777355617946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,256,0.015817777978049386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,256,0.006976889239417181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,128,0.004888888862397936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,768,0.018607111440764535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,128,0.01514488955338796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,6144,128,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,64,0.004603555632962121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,128,0.2975555525885688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,6144,32,0.004530666602982415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,512,0.016629333297411602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,64,0.015142222245534262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,6144,32,0.015119999647140503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,65536,0.27587734328375924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,6144,256,0.3311573399437798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,16384,0.07623200284110175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,65536,0.14029422071244982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,65536,0.6501982476976182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,16384,0.15965688228607178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,12288,0.058528886901007764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,16384,0.038217776351504855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,16384,0.3471715450286865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,12288,0.12416711118486191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,10240,0.04903644323348999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,65536,0.8237102296617297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,12288,0.32434044943915474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,12288,0.029527111185921565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,8192,0.04254577888382805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,10240,0.10512266556421916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,10240,0.025654221574465435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,10240,0.31553954548305935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,7168,0.037751999166276716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,8192,0.021193777521451313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,8192,0.08662399980756973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,8192,0.3252026769849989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,6144,0.03186488813824124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,7168,0.07720800240834554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,7168,0.31919021076626247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,6144,0.06688800123002794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,5120,0.02784622377819485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,6144,0.01793333391348521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,6144,0.3189111020829943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,5120,0.055678221914503306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,5120,0.016453334026866488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,4096,0.02308622168170081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,5120,0.3203199969397651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,4096,0.04627822173966301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,3584,0.021496888664033677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,4096,0.01366311146153344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,4096,0.3207511107126872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,7168,0.01979555520746443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,3584,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,3072,0.01903555625014835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,3584,0.31877777311537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,3072,0.037448889679378934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,2560,0.016345777445369296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,3072,0.011214222345087262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,3072,0.3884302245246039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,2560,0.033484445677863225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,2560,0.3589448928833008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,2560,0.01071200023094813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,2048,0.014021333720948962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,2048,0.02937244375546773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,3584,0.042248000701268516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,1536,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,2048,0.3612062136332194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,2048,0.009358222285906473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,1536,0.0236817780468199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,1536,0.008641777767075432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,1024,0.009457777771684859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,1536,0.342013332578871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,1024,0.019511111908488803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,768,0.008011555506123437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,1024,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,1024,0.36212889353434247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,768,0.017631111873520743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,512,0.006719999843173557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,768,0.007683555285135905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,768,0.34276710616217715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,512,0.015850666496488783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,256,0.005936000082227919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,512,0.007300444775157505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,512,0.34711376825968426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,256,0.015174221661355762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,128,0.005163555757866966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,256,0.007292444507280986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,256,0.34214221106635195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,128,0.014932443698247274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,64,0.004604444321658877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,5120,128,0.006956444846259222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,5120,32,0.006657777975002925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,5120,128,0.3266399966345893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,64,0.014488889111412896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,5120,32,0.014208889669842191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,65536,0.2093635532591078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,16384,0.06091644366582235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,65536,0.13772711488935682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,65536,0.6063244607713487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,16384,0.036751998795403376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,12288,0.04754399922158983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,16384,0.1495822270711263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,65536,0.8116230964660645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,12288,0.028173334068722192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,10240,0.03617777758174472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,12288,0.5416551166110569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,10240,0.09671555625067817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,10240,0.024616888827747766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,16384,0.3374008867475722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,10240,0.47899288601345485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,8192,0.030291554000642564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,8192,0.021008888880411785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,12288,0.11464088492923313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,7168,0.02680533296532101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,8192,0.4315057860480414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,7168,0.32247379091050893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,6144,0.02422222163942125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,6144,0.06131466892030504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,8192,0.08065688610076904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,6144,0.3541413413153754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,5120,0.02100444502300686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,6144,0.017755554782019723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,7168,0.07090755303700765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,5120,0.05100533366203308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,4096,0.018219555417696636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,5120,0.01569333341386583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,7168,0.019161777363883126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,5120,0.3182906574673123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,3584,0.01706755492422316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,4096,0.04305244485537211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,4096,0.013931555880440606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,4096,0.33126666810777455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,3584,0.040038221412234835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,3072,0.015274667077594332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,3584,0.01238666640387641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,3584,0.3203493224249946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,3072,0.035475555393430926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,2560,0.012960000170601739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,3072,0.320525328318278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,2560,0.0314684444003635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,2048,0.011988444460762871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,2560,0.010399999717871347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,2560,0.3256693416171604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,2048,0.0273297776778539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,2048,0.009340444372759925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,1536,0.010088000032636855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,2048,0.32345512178209096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,1536,0.022264000442292955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,1024,0.007358222372002072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,1536,0.00868622213602066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,1536,0.3315102259318034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,1024,0.01852799952030182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,768,0.006533333410819371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,1024,0.007956444389290279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,1024,0.2845519913567437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,3072,0.011212444139851464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,512,0.005624889085690181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,768,0.01683999929163191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,768,0.007647111184067196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,768,0.32215645578172475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,512,0.015485333071814643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,256,0.005200888961553574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,512,0.35540088017781574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,256,0.014868444866604276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,256,0.34486844804551864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,256,0.007038222418891058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,128,0.0048942222363419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,128,0.01442311041884952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,4096,128,0.34024977684020996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,64,0.004521777646409141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,128,0.0069475554757648045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,4096,32,0.004860444615284602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,4096,512,0.007271111011505127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,64,0.014148443937301636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,4096,32,0.01389955480893453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,65536,0.5663928985595703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,65536,0.13358044624328613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,16384,0.04926044411129422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,65536,0.8143493334452311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,65536,0.18314666218227812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,16384,0.13900532987382677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,12288,0.03895466526349386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,16384,0.036151111125946045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,16384,0.4061582353379991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,12288,0.10742577579286362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,10240,0.03420889046457078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,12288,0.027129777603679236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,12288,0.35635821024576825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,10240,0.024315555890401203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,8192,0.028974221812354192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,10240,0.4246515432993571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,8192,0.07537244425879584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,7168,0.024381332927280005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,8192,0.020762667059898376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,8192,0.39489422904120547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,7168,0.06614755259619819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,6144,0.021181333396169875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,7168,0.018611555298169453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,7168,0.3425457742479112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,10240,0.09105333354738022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,6144,0.017266665895779926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,6144,0.057378669579823814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,5120,0.01759288873937395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,6144,0.39436888694763184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,5120,0.04851644568973117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,4096,0.016357334123717416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,5120,0.015216888652907478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,5120,0.38990222083197695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,4096,0.04095110959476895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,4096,0.012770666844315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,3584,0.01440355512830946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,4096,0.37365955776638454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,3584,0.03773777683575948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,3584,0.011791999969217511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,3072,0.013423999978436364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,3584,0.3264426655239529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,2560,0.011511999699804517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,3072,0.03352977832158407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,3072,0.01111288865407308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,2560,0.02954933378431532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,3072,0.35391823450724286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,2048,0.009898666706350114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,2560,0.010082667072614035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,2560,0.32720621426900226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,1536,0.008352888955010308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,2048,0.025387555360794067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,2048,0.009031111167536842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,2048,0.3235235479142931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,1536,0.021329777108298406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,1024,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,1536,0.00870666652917862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,1536,0.3416924476623535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,1024,0.017981333865059745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,768,0.00598933340774642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,1024,0.00794400026400884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,1024,0.3122577667236328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,512,0.0052328887912962176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,768,0.016588444511095684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,768,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,768,0.3292800055609809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,512,0.014917333920796713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,256,0.004545777622196409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,512,0.007326222128338284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,512,0.32693155606587726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,256,0.014129777749379476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,256,0.00700444479783376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,256,0.32580354478624135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,128,0.013827555709415011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,64,0.0038702223036024305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3584,128,0.0069671107663048645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,32,0.0042035554846127825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,64,0.013567110730542077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3584,128,0.34072356753879124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3584,32,0.01347733371787601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,65536,0.16424089007907444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,65536,0.08969066540400188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,16384,0.047761778036753334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,65536,0.5308222240871853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3584,128,0.004233777936961916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,16384,0.12951378027598062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,65536,0.6859555774264865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,16384,0.37459556261698407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,12288,0.0380364457766215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,16384,0.024872889121373493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,12288,0.019620445039537217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,12288,0.10033955838945176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,10240,0.032200889454947576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,12288,0.37784443961249453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,10240,0.08501244253582424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,10240,0.01791911158296797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,8192,0.025662221842341956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,10240,0.36700089772542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,8192,0.0703164471520318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,8192,0.014967110421922473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,7168,0.023053333163261414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,8192,0.4064142174190945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,7168,0.06146577994028727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,6144,0.020096888144810993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,7168,0.014150222142537435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,7168,0.3782924546135797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,6144,0.053194665246539645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,6144,0.012851555314328937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,5120,0.01756800048881107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,6144,0.3987777762942844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,5120,0.04562044474813673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,4096,0.015232000086042615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,5120,0.3477484385172526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,5120,0.011800888511869641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,4096,0.038423998488320246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,3584,0.013658666776286231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,4096,0.010112889111042023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,4096,0.3283662266201443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,3584,0.0350542234049903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,3072,0.01238044434123569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,3584,0.009502222140630087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,3584,0.4042328993479411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,3072,0.03149155444569058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,2560,0.011055111057228513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,3072,0.009069333473841349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,3072,0.33909956614176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,2560,0.02780710988574558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,2048,0.009401777552233802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,2560,0.008383110993438298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,2048,0.023933332827356126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,2048,0.007970666719807519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,1536,0.019655999210145738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,2048,0.47342220942179364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,1536,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,1536,0.3248533407847087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,1024,0.005969777703285217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,1024,0.016837333639462788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,1024,0.007060444189442529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,768,0.005252444495757421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,1024,0.41741334067450625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,1536,0.008028444316652086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,768,0.015612444943851896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,768,0.006919999917348226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,512,0.004578666554556953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,768,0.3650408850775824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,512,0.014665777484575907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,256,0.0041955556306574075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,512,0.3200648890601264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,512,0.006675555474228329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,256,0.014099554883109199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,128,0.003909333298603694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,256,0.006295111030340195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,256,0.3080062336391873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,128,0.013802666631009845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,64,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,2560,0.382767120997111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,3072,128,0.32554933759901256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,64,0.013752000199423896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,3072,32,0.013471110827393003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,65536,0.5168497827317979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,3072,128,0.006240889016124938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,3072,32,0.003564444267087512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,16384,0.042066666815016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,65536,0.08692177799012925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,16384,0.1277893384297689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,65536,0.6770195431179471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,16384,0.3144320117102729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,16384,0.024205333656734888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,12288,0.09832444455888535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,12288,0.019316444794336956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,12288,0.39909688631693524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,10240,0.026562665899594624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,65536,0.14183555708991155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,10240,0.08376000324885051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,8192,0.022566222482257422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,10240,0.33558932940165204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,10240,0.017661333084106445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,12288,0.03459466828240289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,8192,0.014933332800865173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,7168,0.021006221572558086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,8192,0.06823377476798163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,8192,0.31400354703267414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,6144,0.017880888448821176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,7168,0.013624000052611033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,7168,0.06064977910783556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,7168,0.30886843469407826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,6144,0.051868445343441434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,5120,0.015415110521846347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,6144,0.012245333029164208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,6144,0.4066897763146295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,5120,0.044166223870383374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,4096,0.013336000343163809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,5120,0.3359217908647325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,5120,0.01110311100880305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,4096,0.03773244553142124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,3584,0.011864888999197217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,4096,0.010110222631030613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,4096,0.3409288989173041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,3584,0.03437688946723938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,3072,0.010866666833559671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,3584,0.009339555270142024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,3584,0.4148213333553738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,3072,0.030533333619435627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,2560,0.009551111194822524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,3072,0.009002666506502364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,3072,0.35047822528415257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,2560,0.02663555575741662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,2048,0.008415111237102086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,2560,0.008384000096056197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,2048,0.02370044423474206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,2560,0.39683111508687335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,1536,0.007400000260935889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,2048,0.3337048954433865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,2048,0.007727999654081132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,1536,0.019676445258988273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,1024,0.006584888945023219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,1536,0.3490906556447347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,1024,0.016948444975747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,1024,0.3569813304477268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,1024,0.007326222128338284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,768,0.005571555760171678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,768,0.015503999259736804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,768,0.33807555834452313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,512,0.004889777965015835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,768,0.0070017774899800616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,1536,0.007330666813585494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,512,0.014452444182501899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,512,0.006487999939256244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,256,0.004219555606444676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,256,0.013421333498424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,512,0.3998382091522217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,128,0.0038933331767717996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,256,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,256,0.32613333066304523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,64,0.0038951109680864546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,128,0.013455111119482251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2560,128,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2560,128,0.3118720054626465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,64,0.013419555293189155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2560,32,0.01310400002532535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,65536,0.1331928835974799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,65536,0.08502755562464397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,65536,0.4823475413852268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,16384,0.038104888465669416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,16384,0.1190462244881524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,65536,0.669172445933024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,16384,0.33572710884941953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,12288,0.030051555898454454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2560,32,0.003915555361244413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,16384,0.024125332633654278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,10240,0.025154666768179998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,12288,0.09238755702972412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,12288,0.019222221440739103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,12288,0.3126720057593452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,10240,0.07773155636257596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,8192,0.020783111453056335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,10240,0.017288888494173687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,10240,0.32327289051479763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,8192,0.06348266866472033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,7168,0.01796444422668881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,8192,0.014541332920392355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,8192,0.31257955233256024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,6144,0.016136889656384785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,7168,0.05553599860933092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,7168,0.012984889249006907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,7168,0.3205520047081841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,6144,0.04909066690338982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,5120,0.014465777410401238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,6144,0.01219377749496036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,6144,0.32311733563741046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,4096,0.012316444681750404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,5120,0.04145333170890808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,5120,0.011143111520343356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,5120,0.31898400518629283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,3584,0.011067555182509951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,4096,0.035391112168629967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,4096,0.009925333162148794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,4096,0.3206275569068061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,3072,0.010015111002657149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,3584,0.03223022156291538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,3584,0.009304000271691216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,3584,0.32486311594645184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,3072,0.029383109675513372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,3072,0.008730666504965888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,3072,0.3271457884046766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,2560,0.025282666087150574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,2560,0.008379555410808986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,2048,0.007848000360859765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,2560,0.42878399954901797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,2048,0.02202755543920729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,2048,0.00776977754301495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,2048,0.4319297737545437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,2560,0.008817777865462834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,1536,0.018992000155978732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,1536,0.3722853395673964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,1024,0.006037333359320958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,1024,0.016203555795881484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,1536,0.0069573331210348345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,1024,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,768,0.005608888963858287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,1536,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,768,0.015004444453451367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,768,0.006642666541867786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,512,0.004889777965015835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,768,0.42877509858873153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,512,0.014045332868893942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,512,0.006314666734801398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,1024,0.36745421091715497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,256,0.004253333227501975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,512,0.40893867280748153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,256,0.01348977784315745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,128,0.003918222255176968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,256,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,256,0.3513360023498535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,128,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,64,0.0038737778862317405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,2048,128,0.006264888991912206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,2048,32,0.004001777619123459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,2048,128,0.36208266682094997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,64,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,2048,32,0.012747555143303342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,65536,0.11518043941921657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,16384,0.03338133295377096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,65536,0.08218310938941108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,16384,0.10912977986865574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,65536,0.4396417670779758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,12288,0.025627555118666753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,16384,0.3209768931070964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,65536,0.6574639744228786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,12288,0.08293866448932223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,12288,0.018374222848150466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,10240,0.021683555510309007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,12288,0.3227146731482612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,10240,0.07088177733951144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,10240,0.016364443633291457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,8192,0.017806222041447956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,10240,0.38263734181722003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,8192,0.05721955829196506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,7168,0.016025778320100572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,8192,0.014173333843549093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,8192,0.339047114054362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,7168,0.05022222134802076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,6144,0.014499555031458536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,7168,0.013075555364290873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,16384,0.02252977755334642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,7168,0.3607529004414876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,6144,0.044178667995664805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,5120,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,6144,0.32808976703219944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,5120,0.0381040010187361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,4096,0.010845333337783813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,5120,0.011041777829329172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,5120,0.3833377891116672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,4096,0.03267644511328803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,3584,0.009764444496896531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,4096,0.009741333623727163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,4096,0.3322355482313368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,3584,0.029374221960703533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,3072,0.00867733359336853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,3584,0.009177777502271865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,3072,0.026154667139053345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,3584,0.3803199927012126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,6144,0.011552000211344825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,2560,0.007842666572994655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,3072,0.3294088840484619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,3072,0.00868622213602066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,2048,0.00676711110605134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,2560,0.023352000448438857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,2560,0.008385777473449707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,2048,0.020787555310461257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,2560,0.3723742167154948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,1536,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,2048,0.00795999997191959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,2048,0.35591554641723633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,1536,0.017864889568752713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,1024,0.005251555393139522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,1536,0.30762489636739093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,1024,0.015494222442309061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,1024,0.0069777775141927935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,768,0.004576888763242298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,1024,0.353237337536282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,768,0.014459555347760519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,512,0.0042364444169733255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,768,0.006616000086069107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,768,0.38524799876742893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,512,0.013791999883121915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,512,0.006333333336644703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,512,0.3844630983140733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,256,0.0035422220826148987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,1536,0.007319110963079665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,256,0.012818666795889536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,256,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,256,0.37644534640842015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,128,0.012763555679056378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1536,128,0.0063484443558586975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1536,128,0.35837777455647785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,64,0.0033075554917256036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,32,0.0032053333189752367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,64,0.012682666381200155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1536,32,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,65536,0.10263288683361477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1536,128,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,65536,0.43100889523824054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,65536,0.08024089203940497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,16384,0.03347288899951511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,65536,0.6548506418863932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,16384,0.10676711135440403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,12288,0.023159111539522808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,16384,0.373689783944024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,16384,0.02198222279548645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,12288,0.017581333716710407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,10240,0.019205333458052743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,12288,0.3661502202351888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,10240,0.07015022304322985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,10240,0.01591822173860338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,10240,0.3585413297017415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,8192,0.016025778320100572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,8192,0.05721422036488851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,7168,0.014275555809338888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,8192,0.3357440100775824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,8192,0.013632000320487551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,12288,0.08093955781724718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,7168,0.04902844296561348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,6144,0.012845333251688214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,7168,0.31632712152269155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,7168,0.01278133359220293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,5120,0.011328889264000786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,6144,0.011702222128709158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,6144,0.3335386647118463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,5120,0.010757333702511258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,5120,0.03686133358213637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,4096,0.00868977771864997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,5120,0.3270684348212348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,4096,0.030949334303538006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,3584,0.007753777835104201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,4096,0.3403342299991184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,3584,0.028547555208206177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,6144,0.043226665920681424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,3072,0.007364444434642792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,3584,0.0092604441775216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,3072,0.02611733310752445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,3072,0.008632000121805403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,2560,0.006597333484225803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,4096,0.009704888694816166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,3072,0.4412631193796794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,2560,0.023043556345833674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,2048,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,2560,0.008352888955010308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,2560,0.3246471087137858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,3584,0.33032088809543186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,1536,0.0052373334765434265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,2048,0.4128764470418294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,1536,0.01720266706413693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,1536,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,1024,0.004588444613748127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,1536,0.40318044026692706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,1024,0.01520088811715444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,1024,0.0069884442620807225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,2048,0.020598222812016804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,768,0.0041146667467223275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,1024,0.397269328435262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,2048,0.007652444144090016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,768,0.014150222142537435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,512,0.003855111284388436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,768,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,768,0.322613345252143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,512,0.013459555804729462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,256,0.0035217776894569397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,512,0.006623111251327727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,512,0.31356265809800893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,256,0.013021333350075616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,128,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,256,0.006303999986913469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,256,0.3130195670657688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,128,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,64,0.003535110917356279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,1024,32,0.0035208890007601846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,1024,128,0.29471556345621747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,1024,128,0.006312888943486744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,64,0.012411555482281579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,65536,0.0600222216712104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,1024,32,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,16384,0.019567999574873183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,65536,0.08057333363427056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,16384,0.09498577647738987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,65536,0.3917857805887858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,16384,0.33695554733276367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,65536,0.6546186870998806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,16384,0.02161066730817159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,12288,0.01847644481394026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,12288,0.07425066497590807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,10240,0.015628443823920358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,12288,0.4365075429280599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,10240,0.06339199675454034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,10240,0.015474667151769003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,8192,0.01384800010257297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,10240,0.44597509172227645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,8192,0.050407999091678195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,8192,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,8192,0.3716284434000651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,7168,0.012275555895434486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,7168,0.04585777719815572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,7168,0.36306222279866535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,6144,0.01091288857989841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,7168,0.012489777472284106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,6144,0.040567109982172646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,6144,0.011750222080283694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,6144,0.4186506536271837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,5120,0.009418666362762451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,12288,0.017256889078352187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,5120,0.0107351111041175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,5120,0.033618665403789945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,4096,0.008513777620262569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,5120,0.35236620903015137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,4096,0.02849777870708042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,4096,0.009672000176376766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,4096,0.3357528845469157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,3584,0.026755554808510676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,3584,0.33714132838779026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,3584,0.009337777892748514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,3072,0.006871999965773688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,3072,0.024167999625205994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,3072,0.008643555144468943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,2560,0.006330666856633292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,2560,0.021700445148679946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,3072,0.4723466767205133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,2048,0.005630222045713001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,3584,0.007426666716734569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,2560,0.33365244335598415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,2048,0.007635555333561367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,2048,0.019330667124854196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,1536,0.005203555441564984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,2048,0.38789955774943036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,1536,0.016855110724767048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,1024,0.004257777912749184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,1536,0.32583644655015737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,1536,0.007339555356237624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,1024,0.014567999376191033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,768,0.004028444488843282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,1024,0.007043555378913879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,1024,0.3218284447987874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,768,0.01350488927629259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,2560,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,512,0.003605333467324575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,768,0.006599999964237213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,512,0.013167111410035027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,512,0.006245333287451003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,256,0.0032631111227803757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,512,0.3480684492323134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,256,0.012455999851226807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,128,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,256,0.3229831059773763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,128,0.01222400036123064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,128,0.006247111078765657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,128,0.30188266436258954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,64,0.002834666727317704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,768,32,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,64,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,768,32,0.012109333442317115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,65536,0.04464888903829786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,768,768,0.38954843415154355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,65536,0.08053421974182129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,16384,0.01739199956258138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,65536,0.3914942211574978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,768,256,0.00628266649113761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,65536,0.6545111338297526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,16384,0.020696888367335003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,16384,0.09596621990203857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,12288,0.015115555789735583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,16384,0.3356942335764567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,10240,0.013958222336239286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,12288,0.016957332690556843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,10240,0.060512887107001416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,12288,0.39156354798210996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,8192,0.012841777669058906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,10240,0.329128000471327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,10240,0.01549600064754486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,8192,0.050108446015252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,8192,0.013439999686347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,7168,0.01110400011142095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,8192,0.32706133524576825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,7168,0.044176889790429004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,6144,0.010449777874681683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,7168,0.3954666720496283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,12288,0.06989333364698622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,6144,0.03850222296184964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,6144,0.01180266671710544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,5120,0.008654221892356873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,6144,0.4088791211446126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,5120,0.033589333295822144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,5120,0.01072177787621816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,4096,0.007799111306667328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,5120,0.42026223076714414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,7168,0.012431999875439538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,4096,0.028718223174413044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,4096,0.009684444301658207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,4096,0.3968782160017226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,3584,0.026662222213215295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,3584,0.009054222040706212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,3072,0.006931555353932911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,3584,0.41565778520372176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,3072,0.02372088862790002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,3072,0.008527999950779809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,2560,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,3072,0.3993137677510579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,2560,0.021319111188252766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,2560,0.008361777497662438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,2560,0.41195376714070636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,2048,0.005466666486528184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,2048,0.019315555691719055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,2048,0.35704800817701554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,1536,0.00489777781897121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,2048,0.007739555504586961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,3584,0.007256000406212277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,1536,0.016545777519543965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,1024,0.004293333325121137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,1536,0.007303111255168915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,1536,0.38175021277533633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,1024,0.014478221535682678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,768,0.004228444563017951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,1024,0.007000000112586551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,768,0.013412444127930535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,1024,0.319997337129381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,768,0.006695999867386288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,512,0.012796444197495779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,512,0.006315555423498154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,256,0.003228444399105178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,512,0.43798666530185276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,256,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,256,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,128,0.0032622222271230486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,256,0.39220621850755477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,512,0.003588444242874781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,128,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,64,0.003160888950030009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,512,128,0.006227555374304454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,768,0.36339733335706925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,512,32,0.0032151111712058387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,65536,0.0329440004295773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,64,0.012078222301271228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,512,128,0.3364951080746121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,512,32,0.011824000212881299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,16384,0.013009777499569787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,65536,0.07928711175918579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,65536,0.3898915449778239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,16384,0.09312888648774888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,12288,0.010837333069907295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,65536,0.6523911158243815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,16384,0.020632889535692003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,16384,0.3444080087873671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,10240,0.009697777529557547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,12288,0.016934222645229764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,12288,0.07265066438251071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,12288,0.32712801297505695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,8192,0.009278222090668148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,10240,0.01519466605451372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,10240,0.33054934607611763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,8192,0.048750221729278564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,7168,0.008623999853928884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,8192,0.013226666384273104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,8192,0.39032800992329914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,7168,0.04412177867359585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,6144,0.008356444537639618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,7168,0.32340534528096515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,7168,0.012446222205956778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,6144,0.03804444604449802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,5120,0.008678221868144141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,6144,0.01144177797767851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,6144,0.31836533546447754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,5120,0.03242933419015672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,4096,0.007618666523032718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,5120,0.0107342220014996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,5120,0.3992284403906928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,4096,0.3160773383246528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,4096,0.009675555759006077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,3584,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,3584,0.026143999563323125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,3584,0.3184000121222602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,3072,0.006587555425034628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,3584,0.009036444127559662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,10240,0.0609075559510125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,3072,0.023676445086797077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,2560,0.005986666513813867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,3072,0.008360889222886827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,4096,0.028191109498341877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,2560,0.02134311033619775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,2560,0.3566577699449327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,2560,0.008191111187140146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,2048,0.005251555393139522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,2048,0.018995554910765756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,2048,0.007673777639865875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,2048,0.3579111099243164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,1536,0.004860444615284602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,1536,0.016180444094869826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,1536,0.3618568844265408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,3072,0.38917867342631024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,1024,0.004311111238267687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,1024,0.014515555567211576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,1024,0.35513777203030056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,768,0.0038373333712418876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,1024,0.006994666324721442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,768,0.01349066694577535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,768,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,512,0.0037946667936113146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,768,0.4293617672390408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,1536,0.007335111498832703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,512,0.01239822225438224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,256,0.0032106666929192017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,512,0.0063075555695427795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,512,0.3688231044345432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,256,0.01201600001917945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,128,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,256,0.006263999889294307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,256,0.353420443005032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,128,0.011707555916574268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,64,0.0031511110977994073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,256,128,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,256,32,0.0028888889484935333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,64,0.011763555308183035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,256,128,0.37168267038133407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,65536,0.02513866623242696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,256,32,0.01162755572133594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,65536,0.07779288954204984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,16384,0.011446221835083432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,65536,0.6532533433702257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,16384,0.09491911199357773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,12288,0.010004444254769219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,16384,0.020643555455737643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,12288,0.0718053314420912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,16384,0.4812302059597439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,10240,0.009450666606426239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,12288,0.32275022400750053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,12288,0.01684888866212633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,10240,0.015226667126019796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,10240,0.060699555608961314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,8192,0.008646222452322641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,65536,0.3898444440629747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,10240,0.3266293207804362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,7168,0.008311111066076491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,8192,0.04761244522200691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,8192,0.013370667066839008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,7168,0.04253866606288486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,8192,0.3304702175988091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,6144,0.007662222617202335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,7168,0.012470222181744046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,6144,0.037909332248899676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,7168,0.37239644262525773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,5120,0.008008889026112026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,6144,0.011518222590287527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,6144,0.3232195642259386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,4096,0.007633777956167857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,5120,0.03265155686272515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,5120,0.010974222587214576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,5120,0.32166756523980033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,4096,0.02808977829085456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,3584,0.006966222491529252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,4096,0.009718221922715506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,4096,0.35298755433824325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,3072,0.006603555546866522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,3584,0.00908799966176351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,3584,0.31807467672559947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,3072,0.02369600037733714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,3072,0.008343111309740279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,2560,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,3072,0.37976622581481934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,2560,0.02103999919361538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,2048,0.005256888767083486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,2560,0.33079732788933647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,2560,0.008216888540320927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,2048,0.018932445181740653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,1536,0.004589333302444882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,2048,0.007649777664078607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,2048,0.33252178298102486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,1536,0.016173332929611206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,1024,0.004184000194072723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,1536,0.007273777491516537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,1024,0.014214222629865011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,1536,0.3944213390350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,768,0.003912000192536248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,1024,0.0069146665434042616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,1024,0.32573334376017254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,768,0.013416888813177744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,512,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,768,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,768,0.3251111242506239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,512,0.012481778032249875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,3584,0.02610133257177141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,512,0.36449511845906574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,256,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,512,0.006325333482689328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,256,0.012135110795497894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,128,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,256,0.34789334403143984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,128,0.012111110819710625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,128,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,64,0.002834666727317704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,64,0.011773332953453064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,128,32,0.0028702221396896574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,768,128,256,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,128,32,0.011392888923486074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,65536,0.022072000635994807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,16384,0.009783999787436591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,65536,0.3896346622043186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,16384,0.09506844149695502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,768,128,128,0.3518986701965332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,12288,0.008365333080291748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,10240,0.009374221993817223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,8192,0.008411555654472774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,12288,0.07224533292982313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,10240,0.05960622098710802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,8192,0.04729955395062765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,6144,0.007682667010360294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,7168,0.042656001117494374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,5120,0.00832088871134652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,6144,0.037147555086347796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,4096,0.007609777980380588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,5120,0.032300445768568255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,3584,0.007287110719415877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,4096,0.027760000692473516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,3072,0.006917333437336816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,3584,0.026076444321208533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,2560,0.006603555546866522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,3072,0.023404444257418316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,2048,0.004926222066084544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,7168,0.008348444269763099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,2560,0.021272889441914026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,1536,0.004569777597983678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,2048,0.019016888406541612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,1024,0.003958222352796131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,768,0.003601777884695265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,1536,0.016502221425374348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,1024,0.014197332991494073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,256,0.0032142222755485107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,768,0.013225778109497495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,512,0.0127368892232577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,128,0.002903999967707528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,256,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,64,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,32,0.0029351111087534162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,128,0.011726222104496427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,65536,0.025784000754356384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,64,0.011755555868148804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,16384,0.010664888554149203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,64,32,0.011536888778209686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,12288,0.00904088881280687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,16384,0.09420977698432074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,10240,0.008602666358153025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,65536,0.3897973431481256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,8192,0.00830844458606508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,12288,0.07090044683880277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,7168,0.008001777860853408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,64,512,0.0032684444967243406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,10240,0.06025866667429606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,6144,0.007705777883529663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,5120,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,8192,0.04688444402482775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,4096,0.007296000089910295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,7168,0.04140888982348972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,6144,0.03701688845952352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,3584,0.006992888947327931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,3072,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,5120,0.03241066800223456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,4096,0.028218666712443035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,2560,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,3584,0.02608266638384925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,2048,0.00535733335547977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,1536,0.004569777597983678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,3072,0.023059555225902136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,2560,0.02128000060717265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,1024,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,2048,0.01896711190541585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,768,0.0035546666218174826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,512,0.0035617777870761026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,1536,0.01630311045381758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,1024,0.014549333188268872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,256,0.0032479998966058097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,768,0.01310755560795466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,128,0.0028968888024489083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,512,0.0124177775449223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,64,0.0029271110478374693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,768,32,32,0.002944888960984018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,256,0.012098666694429187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,128,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,64,0.011606222225560082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,768,32,32,0.011500444677140979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,16384,0.5930497911241319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,16384,0.6960764461093478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,12288,0.4566684299045139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,16384,0.21319466167026094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,12288,0.5127333535088433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,16384,0.6623697810702854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,10240,0.41349511676364475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,12288,0.5179591178894043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,10240,0.43466666009691024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,10240,0.13319910897148982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,8192,0.26615111033121747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,10240,0.444431119494968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,8192,0.34511200586954754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,7168,0.23421955108642578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,8192,0.11346844832102458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,12288,0.15909600257873535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,8192,0.36321245299445254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,7168,0.3016177813212077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,7168,0.09894844558503892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,6144,0.21908622317843965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,7168,0.33719200558132595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,6144,0.25952355066935223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,6144,0.08685333198971218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,5120,0.17848177750905356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,6144,0.32251021597120494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,5120,0.21952088673909506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,5120,0.07843822240829468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,4096,0.14947466055552164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,5120,0.3270337846544054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,4096,0.06317955255508423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,4096,0.17957866191864014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,3584,0.12603021992577448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,3584,0.1603448920779758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,4096,0.3342408869001601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,3584,0.05493155452940199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,3072,0.10671733485327826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,3072,0.14147022035386828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,3584,0.4229244391123454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,2560,0.09451822439829509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,3072,0.326928006278144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,2560,0.12187466356489395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,2048,0.07662755250930786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,2560,0.33236267831590444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,2048,0.09939999712838067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,2048,0.32055287890964085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,2048,0.03906666570239597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,3072,0.05020977722273933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,1536,0.06183466646406385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,1536,0.08040355311499701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,1536,0.03315022256639268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,1536,0.3378560013241238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,2560,0.04534133275349935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,1024,0.06345155504014757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,1024,0.02555466691652934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,1024,0.32262134552001953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,768,0.05199911197026571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,768,0.32743467224968803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,768,0.022416889667510986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,512,0.028255999088287354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,512,0.045828445090187915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,1024,0.042782220575544566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,256,0.020301333732075162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,512,0.32642843988206655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,512,0.019668444991111755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,768,0.03705866800414191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,256,0.04045066568586562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,128,0.01848088867134518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,256,0.3269768820868598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,128,0.037639111280441284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,64,0.014923555983437432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,128,0.017749332719379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,65536,32,0.015807999504937064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,64,0.03706044289800856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,65536,32,0.03703466720051236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,65536,256,0.017902221944597032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,65536,0.55913331773546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,65536,128,0.31288621160719127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,65536,0.8572888904147677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,16384,0.1539964410993788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,16384,0.2214284472995334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,65536,0.24043732219272188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,12288,0.11606310473548041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,16384,0.0665795538160536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,16384,0.33551732699076336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,12288,0.16480533281962076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,10240,0.10461955600314671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,12288,0.051757332351472646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,12288,0.3287546634674072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,10240,0.14090666506025526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,10240,0.045316444502936475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,8192,0.08398844136132134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,10240,0.3230053318871392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,8192,0.11586577362484402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,8192,0.03801600138346354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,7168,0.07627199755774604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,8192,0.4137680000729031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,7168,0.10291022062301636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,6144,0.06227999925613403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,7168,0.0332897769080268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,6144,0.08971999751196967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,6144,0.4632115364074707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,65536,0.9413262473212348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,5120,0.05359111229578654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,6144,0.029256890217463177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,5120,0.07699200179841784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,7168,0.3025884363386366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,4096,0.045528001255459256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,5120,0.4095279905531142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,5120,0.02508977717823452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,4096,0.0631964471605089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,4096,0.020963556236690942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,3584,0.0399528874291314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,4096,0.38955020904541016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,3584,0.05689777930577596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,3072,0.03455288873778449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,3584,0.34600623448689777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,3584,0.019155555301242404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,3072,0.0510791109667884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,2560,0.02935911218325297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,3072,0.017641777793566387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,3072,0.3477457894219293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,2560,0.04443466663360596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,2048,0.024109333753585815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,2560,0.413375112745497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,2048,0.03596533338228861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,2048,0.014096889230940076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,1536,0.019851555426915485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,2048,0.4231119950612386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,1536,0.030163556337356567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,1536,0.011857777833938599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,1024,0.014568888478808932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,1536,0.40739112430148655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,2560,0.015969778100649517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,1024,0.025038222471872967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,1024,0.00998755544424057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,1024,0.39719910091824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,768,0.0222382214334276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,512,0.009894222021102905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,512,0.019924443629052904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,768,0.37705156538221574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,512,0.008997333546479544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,512,0.32761510213216144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,256,0.009363555245929295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,256,0.018584888842370775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,768,0.012236444486512078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,128,0.006537777682145436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,256,0.008447999755541483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,256,0.3880000114440918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,128,0.017872000734011333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,64,0.0076142218377855085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,128,0.008472888833946651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,16384,128,0.300311115052965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,16384,32,0.007997333175606197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,64,0.01761244402991401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,16384,32,0.017851556340853374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,65536,0.42184265454610187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,65536,0.707449754079183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,16384,0.11532710658179389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,65536,0.1892951064639621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,16384,768,0.009463999834325578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,16384,0.18093600538041857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,12288,0.08761155605316162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,16384,0.053257776631249316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,16384,0.388321770562066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,12288,0.1395297712749905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,10240,0.0736373331811693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,12288,0.041089776489469744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,65536,0.804707580142551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,12288,0.42030755678812665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,8192,0.06272977590560913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,10240,0.11756622791290283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,10240,0.03646488984425863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,8192,0.09824177953932021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,10240,0.38013778792487246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,8192,0.029751999510659113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,7168,0.05573155482610067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,8192,0.3212391005622016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,6144,0.04901955525080363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,7168,0.026345777842733595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,7168,0.3287999894883898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,6144,0.07562844620810615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,5120,0.04140355520778232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,6144,0.023204444183243647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,6144,0.36539732085333926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,5120,0.06628800100750394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,4096,0.033794667985704206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,5120,0.02050488856103685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,5120,0.3226666715410021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,4096,0.05305511090490553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,4096,0.017456000049908955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,4096,0.3593182298872206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,7168,0.08589155806435479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,3584,0.04759644468625387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,3072,0.026877333720525105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,3584,0.01612444387541877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,3584,0.33408710691663956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,3072,0.04260355565283033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,2560,0.02286311156219906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,3072,0.014799111419253878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,2560,0.03825689024395413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,3072,0.40045777956644696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,3584,0.029314666986465454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,2560,0.36362398995293516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,2560,0.01368533323208491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,2048,0.01906399925549825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,2048,0.03185688787036472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,2048,0.012089778151777057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,1536,0.01571377780702379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,2048,0.3907253212398953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,1536,0.026692445079485577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,1024,0.01184177812602785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,1536,0.3645724455515544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,1024,0.02241244415442149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,1024,0.008653333617581261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,768,0.010262221925788457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,768,0.01958577831586202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,1536,0.010319111247857412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,768,0.4366062217288547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,512,0.008546666966544257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,768,0.008338666624493068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,512,0.01792622274822659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,512,0.0079644446571668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,256,0.006728888799746831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,512,0.4156844351026747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,256,0.01663822266790602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,256,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,128,0.006578666468461354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,128,0.016191111670600045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,1024,0.3838684558868408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,12288,128,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,64,0.006239999913507038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,128,0.34118933147854275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,12288,32,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,64,0.016175111134847004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,12288,32,0.015812445018026564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,12288,256,0.42976623111301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,65536,0.39629422293768984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,16384,0.10898933145735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,65536,0.6308008829752604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,65536,0.18170489205254448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,16384,0.1741724411646525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,12288,0.0871395534939236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,16384,0.3287893401251899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,12288,0.12585155169169107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,12288,0.04030044542418586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,65536,0.7991395526462131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,10240,0.06590044498443604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,12288,0.3346693250868056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,10240,0.03343288766013251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,10240,0.10605421993467544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,8192,0.051855110459857516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,10240,0.31016267670525444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,16384,0.050627556112077504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,7168,0.049365331729253135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,8192,0.08597777949439155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,8192,0.02850488821665446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,8192,0.32736534542507595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,7168,0.025269332859251235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,6144,0.04089777668317159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,7168,0.36262043317159015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,6144,0.06831022103627522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,5120,0.03613244493802389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,6144,0.3206640084584554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,6144,0.022250667214393616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,5120,0.05825155311160617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,5120,0.01978933314482371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,4096,0.03038755390379164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,5120,0.31673423449198407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,4096,0.047022223472595215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,4096,0.016906667086813185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,3584,0.025751110580232408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,4096,0.33526844448513454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,7168,0.07648355431026883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,3072,0.023640000157886084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,3584,0.04264622264438205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,3584,0.34124000867207843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,3072,0.038479109605153404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,2560,0.020418667131000094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,3072,0.3290933238135444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,3072,0.014448889427714877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,2560,0.03460355599721273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,2560,0.013332444760534497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,2048,0.017264889346228707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,2560,0.31832265853881836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,2048,0.02903555499182807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,1536,0.01260088880856832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,2048,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,2048,0.3852710988786485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,1536,0.024737776981459722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,3584,0.015623999966515435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,1024,0.009848000274764167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,1536,0.32059645652770996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,1536,0.009707555174827576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,1024,0.020595555504163105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,1024,0.008664000365469191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,1024,0.33164000511169434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,768,0.018336888816621568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,512,0.007015110717879401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,768,0.008004444340864817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,768,0.3181066777971056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,512,0.01689511040846507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,512,0.007978666987684038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,256,0.006367110957702001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,512,0.37896354993184406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,256,0.016169778174824186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,768,0.009217778013812171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,128,0.005646222167544895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,256,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,128,0.015620445211728414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,128,0.341048002243042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,64,0.005244444641802046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,10240,128,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,10240,32,0.005316444569163852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,64,0.015576000014940897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,10240,32,0.015114666687117683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,65536,0.29066043429904515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,10240,256,0.3575590981377496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,65536,0.5520719952053493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,65536,0.1369431151284112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,16384,0.08139822218153211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,16384,0.14217066764831543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,16384,0.039048890272776283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,12288,0.06014221906661987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,65536,0.6617502106560601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,12288,0.1080462204085456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,12288,0.3510506682925754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,12288,0.03026755650838216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,10240,0.054647111230426364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,10240,0.09219111336602105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,16384,0.3714951144324408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,10240,0.3435697820451524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,8192,0.044176889790429004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,10240,0.02611911131276025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,7168,0.037477334340413414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,8192,0.021873777111371357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,8192,0.07518666982650757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,7168,0.06658133533265856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,8192,0.3220275508032905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,6144,0.03288177649180094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,7168,0.019354666272799175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,6144,0.05950578053792318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,7168,0.3886115550994873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,5120,0.0288426677385966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,6144,0.017643555998802185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,6144,0.3302266597747803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,4096,0.02367555598417918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,5120,0.05052177773581611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,5120,0.015964445140626695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,5120,0.33132889535692006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,4096,0.04137600130505032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,3584,0.02144711050722334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,4096,0.013947554760509066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,3584,0.037873778078291155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,3072,0.019104000594880845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,3584,0.012964444855848948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,3584,0.32419554392496747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,3072,0.034113778008355036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,2560,0.016984888248973422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,3072,0.3680364555782742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,2560,0.030817776918411255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,2560,0.010803555448849997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,2048,0.015132443772421943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,2560,0.3409173223707411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,2048,0.02614044480853611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,4096,0.36056799358791775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,1536,0.01236977759334776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,2048,0.009399111072222391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,2048,0.33064622349209255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,1536,0.022704000274340313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,3072,0.011983999775515662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,1536,0.00832711077398724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,1536,0.33183378643459743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,1024,0.018906666172875296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,768,0.008355555435021719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,1024,0.007624000310897827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,1024,0.33419733577304417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,768,0.01718933383623759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,512,0.00629333323902554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,768,0.3393875492943658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,512,0.016022221909628976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,512,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,256,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,512,0.331404447555542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,1024,0.00978488889005449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,256,0.015537778536478678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,128,0.004889777965015835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,256,0.006949333681000604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,256,0.3198622332678901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,128,0.015174221661355762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,128,0.006717333363162146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,64,0.004565333326657613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,8192,768,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,8192,128,0.36260355843438047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,64,0.01478488908873664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,8192,32,0.014561777313550314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,65536,0.25980888472663033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,65536,0.5112053553263346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,16384,0.06997955507702298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,65536,0.13296622700161403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,8192,32,0.006401777681377199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,16384,0.130657778845893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,16384,0.03854399919509888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,65536,0.6514737870958116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,12288,0.05387644304169548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,12288,0.10001777940326267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,12288,0.3700444433424208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,12288,0.028971556160185073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,10240,0.048228445980283946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,10240,0.0855555534362793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,10240,0.36403467920091415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,16384,0.42030933168199325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,10240,0.025182222326596577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,8192,0.04008977942996555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,8192,0.02158222264713711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,8192,0.07008710834715101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,7168,0.03339022066858079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,8192,0.36729688114590114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,7168,0.06204444169998169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,7168,0.019095111224386428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,6144,0.03211999932924906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,7168,0.36000267664591473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,6144,0.05372711022694906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,6144,0.0177839994430542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,5120,0.0251519994603263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,6144,0.4083724551730686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,5120,0.045927110645506114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,4096,0.022783999641736347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,5120,0.016012445092201233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,5120,0.35878756311204696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,4096,0.03902755512131585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,4096,0.013880000346236758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,4096,0.37723467085096574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,3584,0.03543377916018168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,3584,0.01258399999803967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,3072,0.018234666850831773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,3584,0.37881867090861004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,3072,0.03226577904489305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,3072,0.011588444312413534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,2560,0.016626665989557903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,3072,0.3354826768239339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,3584,0.020455999506844413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,2048,0.01293333371480306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,2560,0.010470222267839642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,2048,0.024700444605615404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,2560,0.46659109327528214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,1536,0.011447110937701331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,2048,0.009302222066455418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,2048,0.37376801172892254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,1536,0.021692444880803425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,1536,0.008381333616044786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,1024,0.00925155563486947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,1536,0.33356889088948566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,1024,0.01757777730623881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,1024,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,768,0.00792711145348019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,2560,0.02908710969818963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,1024,0.3954186704423692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,768,0.016545777519543965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,512,0.006292444550328785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,768,0.007296000089910295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,512,0.015569777952300178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,768,0.37393511666191953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,256,0.004960888789759742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,512,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,512,0.3697422345479329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,256,0.015185778339703878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,128,0.00444444434510337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,256,0.3278435601128472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,128,0.014550222290886773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,64,0.0042204442951414315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,128,0.006700444552633498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,7168,32,0.004208000169859992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,64,0.014535999960369535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,7168,32,0.01447111037042406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,65536,0.1983893314997355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,7168,256,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,65536,0.4913315773010254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,16384,0.05484799875153435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,7168,128,0.3332666556040446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,16384,0.12653422355651855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,16384,0.037083556254704796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,65536,0.6526995764838325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,16384,0.3451022307078044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,12288,0.09909688764148289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,12288,0.3446764416164822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,10240,0.04159288936191135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,12288,0.02863022353914049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,65536,0.1289608875910441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,10240,0.08482578065660264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,8192,0.034188446071412824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,10240,0.024124445186720953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,10240,0.3869955539703369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,12288,0.04293955696953667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,8192,0.06987288925382826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,7168,0.027391110857327778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,8192,0.02109955582353804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,8192,0.32637778917948407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,6144,0.026551999979548987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,7168,0.06062489085727268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,7168,0.019205333458052743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,7168,0.33005065388149685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,6144,0.05209866828388638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,5120,0.02216888964176178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,6144,0.35342754258049863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,6144,0.01742044422361586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,5120,0.04537777768241035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,5120,0.015610666738616096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,4096,0.01899466746383243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,5120,0.34188800387912327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,4096,0.038057777616712786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,3584,0.015764445066452026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,4096,0.013713777893119387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,4096,0.373872889412774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,3584,0.0350213348865509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,3584,0.01239111108912362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,3072,0.015434667468070984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,3584,0.35700445704989964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,3072,0.03196622265709771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,2560,0.014498665928840637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,3072,0.011415999796655444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,3072,0.367304007212321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,2560,0.028786665863460962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,2048,0.01200355589389801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,2560,0.010309333602587381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,2560,0.3624675538804796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,2048,0.023932443724738225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,1536,0.00977422214216656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,2048,0.00904444439543618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,2048,0.34984800550672746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,1536,0.020255999432669747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,1024,0.008043555749787224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,1536,0.008184000021881526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,1536,0.35659999317593044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,1024,0.017234666479958426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,768,0.006611555400821898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,1024,0.007622222105662028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,1024,0.3514355553521051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,768,0.016184000505341422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,512,0.005576888720194499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,768,0.007292444507280986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,768,0.3426453272501628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,512,0.01479466590616438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,256,0.004945777770545748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,512,0.006948444578382704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,512,0.3468622101677789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,256,0.014462222655614218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,128,0.004922666483455234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,256,0.006638222270541721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,256,0.29972532060411244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,64,0.0045422220395671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,64,0.014123555686738757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,6144,128,0.31529421276516384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,6144,32,0.004893333133724001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,32,0.01388088862101237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,65536,0.44917954338921445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,6144,128,0.014167999227841696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,6144,128,0.006815111057625876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,16384,0.054096887509028115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,65536,0.12658488750457764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,65536,0.17828533384535047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,16384,0.116138670179579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,16384,0.03540800015131632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,12288,0.04184533490075005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,16384,0.3329511218600803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,65536,0.6566986507839626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,10240,0.03569511241383023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,12288,0.08945155806011623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,12288,0.025938666529125635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,12288,0.3923315472073025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,10240,0.023108444280094568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,10240,0.33937777413262266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,8192,0.06256177690294054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,8192,0.019984000258975558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,8192,0.34038933118184406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,7168,0.02584533393383026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,7168,0.05492711067199707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,7168,0.3396880096859402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,7168,0.01831377711560991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,10240,0.07780355877346463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,6144,0.04850133260091146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,6144,0.016882666283183627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,6144,0.31646045049031574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,5120,0.01991555591424306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,5120,0.04192355606291029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,5120,0.33612267176310223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,5120,0.014711999230914645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,4096,0.01684888866212633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,6144,0.023007111416922674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,4096,0.035700443718168474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,8192,0.02988622254795498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,4096,0.012770666844315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,4096,0.3344702190823025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,3584,0.03271466824743483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,3584,0.012058667010731168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,3072,0.013728888498412238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,3584,0.3305404451158312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,3072,0.030212445391549006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,2560,0.012456888953844706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,3072,0.01076533314254549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,3072,0.38925332493252224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,2560,0.026708443959554035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,2560,0.35869601037767196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,2048,0.010493333141009012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,3584,0.014843554960356818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,2048,0.022167111436525982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,2048,0.3776222334967719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,2048,0.00868711123863856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,1536,0.009242666264375051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,1536,0.019296889503796894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,1536,0.3704675568474664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,1536,0.007976888782448238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,1024,0.00700711127784517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,2560,0.009695111049546136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,1024,0.016546666622161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,768,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,1024,0.007600888609886169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,1024,0.3482666545444065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,768,0.015549333559142219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,512,0.005296000176005893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,768,0.007259555160999298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,768,0.3483724329206679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,512,0.014446222119861178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,256,0.004892444445027245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,512,0.006979555719428592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,512,0.34845956166585285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,256,0.013773333695199756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,128,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,256,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,256,0.35052000151740176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,128,0.013532444834709167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,64,0.0038933331767717996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,5120,128,0.006614222294754452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,5120,128,0.3145502143436008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,5120,32,0.0041946665280395085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,64,0.013437333206335703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,5120,32,0.013549333645237817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,65536,0.4038577874501546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,16384,0.04639910989337497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,65536,0.08059110906389025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,16384,0.10452089044782852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,65536,0.5187253422207302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,12288,0.03791200121243795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,16384,0.023775110642115276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,65536,0.16843911012013754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,12288,0.08254844612545438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,10240,0.03276177909639146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,16384,0.5256097581651475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,12288,0.01906488835811615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,12288,0.324855989880032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,8192,0.025750221477614507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,10240,0.06950044631958008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,10240,0.017071111334694754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,8192,0.05579644441604614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,8192,0.015051555302408008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,7168,0.02275644408331977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,10240,0.4492204454210069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,7168,0.04889688889185587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,7168,0.013550221920013428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,6144,0.02052088909678989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,7168,0.3262275589836968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,6144,0.04336888922585381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,6144,0.012463111016485425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,5120,0.01799199978510539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,6144,0.33859554926554364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,5120,0.0383840004603068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,5120,0.011026666396194033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,4096,0.014624888698259989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,5120,0.38390223185221356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,4096,0.03288622366057502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,3584,0.013481777575280933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,8192,0.3145413398742676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,4096,0.3350364367167155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,3584,0.030234667989942763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,3072,0.012159999873903064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,3584,0.009012444151772393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,3584,0.30814756287468803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,3072,0.02784622377819485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,3072,0.008954666554927826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,2560,0.010942222343550788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,3072,0.3212133248647054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,2560,0.02551555633544922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,2048,0.009711110757456886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,2560,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,2560,0.31997778680589467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,2048,0.020676443974177044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,1536,0.007445333732499017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,2048,0.0076142218377855085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,2048,0.3196595509847005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,4096,0.009678222239017487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,1024,0.006054222169849608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,1536,0.018412444326612685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,1536,0.0069368887278768755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,1536,0.3205146789550781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,1024,0.01585155559910668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,768,0.005244444641802046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,1024,0.006666666517655055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,1024,0.310714668697781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,512,0.004881777697139316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,768,0.014848889576064216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,768,0.3330559995439318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,512,0.014470222923490735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,512,0.006280888699822956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,256,0.0041857775714662345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,512,0.42373956574334043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,256,0.013838222457302941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,256,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,256,0.40100354618496364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,768,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,128,0.013274666335847644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,4096,128,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,64,0.00387999994887246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,4096,128,0.35892176628112793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,32,0.003630222131808599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,64,0.013166222307417126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,4096,32,0.012811555630630918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,65536,0.15338667233784994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,4096,128,0.004212444441186057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,65536,0.07900177770190768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,16384,0.04397866792149014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,16384,0.09615200095706516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,65536,0.5139590899149576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,16384,0.023200000325838726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,12288,0.03423822257253859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,16384,0.37460711267259383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,12288,0.07488266626993816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,65536,0.3706328868865967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,12288,0.34990843137105304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,10240,0.028468443287743464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,12288,0.01863466699918111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,10240,0.06352088848749797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,8192,0.023644444015291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,10240,0.016744888491100736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,10240,0.3426915539635553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,8192,0.05075110991795858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,7168,0.021070222059885662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,8192,0.3897537655300564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,7168,0.04537511203024122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,7168,0.32835290167066783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,7168,0.013205333716339536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,6144,0.04032355546951294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,6144,0.012176888684431711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,6144,0.4043413268195258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,5120,0.015790222419632804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,8192,0.01460088955031501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,5120,0.3353031211429172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,4096,0.013754666679435305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,5120,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,6144,0.01865688959757487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,4096,0.030574222405751545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,3584,0.012561777399645912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,4096,0.32705865965949166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,4096,0.0094373333785269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,3584,0.028233776489893597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,3072,0.011136888629860349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,3584,0.008974221845467886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,3584,0.3229831059773763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,3072,0.025418667329682246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,2560,0.009919111099508073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,3072,0.0083155557513237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,2560,0.02274666726589203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,3072,0.3879235585530599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,2048,0.008068444828192392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,2560,0.33168800671895343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,2048,0.019855111837387085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,2048,0.007616000043021307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,1536,0.00674933319290479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,1536,0.017280000779363845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,5120,0.035544888840781316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,1536,0.0070773329999711775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,2560,0.00795555528667238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,1024,0.005579555614127054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,1536,0.35681867599487305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,1024,0.015493333339691162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,1024,0.00665244460105896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,768,0.005233777893914117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,1024,0.3806648784213596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,768,0.014458666245142618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,512,0.004519999855094486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,768,0.32142755720350474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,768,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,512,0.014103111293580798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,256,0.0038906666967603895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,512,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,512,0.3157546785142687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,256,0.013418667018413544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,128,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,256,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,128,0.013140444126394061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,256,0.39254577954610187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,64,0.0035946667194366455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3584,128,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3584,32,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,64,0.012814222110642327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3584,32,0.012910222013791403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,65536,0.14731733004252115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,65536,0.36282934082878965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,65536,0.0769395563337538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,65536,0.5103235774570042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,16384,0.04240622123082479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,128,0.2965262201097277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3584,2048,0.32685777876112193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,16384,0.09451022413041855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,16384,0.022639999787012737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,12288,0.03268355462286208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,16384,0.3731733428107367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,12288,0.07253422339757283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,12288,0.018584888842370775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,10240,0.026755554808510676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,12288,0.36032888624403214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,10240,0.0618399977684021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,8192,0.023176888624827068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,10240,0.016540444559521146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,10240,0.36496443218655056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,8192,0.050028443336486816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,7168,0.02036266691154904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,8192,0.0140364451540841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,8192,0.3808755609724257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,7168,0.04401510953903198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,6144,0.017954667409261067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,7168,0.34860889116923016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,7168,0.012802666260136498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,6144,0.03976000017589993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,6144,0.011787555283970304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,6144,0.3556364377339681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,5120,0.035339554150899254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,5120,0.34529956181844074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,4096,0.013267555170589022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,4096,0.02988622254795498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,5120,0.015437333120240105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,4096,0.36560267872280544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,4096,0.009687999884287516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,3584,0.01202044470442666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,3584,0.02772533396879832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,5120,0.010440000229411654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,3072,0.010827555424637265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,3584,0.3338008986579047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,3584,0.008976889153321585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,3072,0.025118221839269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,3072,0.008358221915033128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,2560,0.009831111464235518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,3072,0.32351377275254994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,2560,0.02197955548763275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,2048,0.008467555873923832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,2560,0.008035555481910706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,2560,0.38905244403415257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,2048,0.01925244430700938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,1536,0.006745777610275481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,2048,0.00756444450881746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,2048,0.32090311580234104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,1536,0.01720622181892395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,1536,0.007009777757856581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,1024,0.015164444843928019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,1536,0.38359912236531574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,768,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,1024,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,768,0.014466666513019137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,1024,0.41919199625651044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,512,0.004954666727119022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,768,0.3265795442793104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,768,0.006405333264006509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,512,0.013798221945762634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,256,0.004233777936961916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,512,0.006239111224810283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,1024,0.005722666780153911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,512,0.3253493309020996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,128,0.003944000022278892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,256,0.013405332962671915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,256,0.005910222315125995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,128,0.013030221892727746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,256,0.3705457846323649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,64,0.003867555409669876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,3072,128,0.0058959999846087555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,3072,32,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,3072,128,0.3661235438452826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,65536,0.11418844593895806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,64,0.012795555922720166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,3072,32,0.01239466667175293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,16384,0.03201422095298767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,65536,0.07525066534678142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,16384,0.09244800276226467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,65536,0.35772177908155656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,12288,0.025444444682863023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,16384,0.022135110365019903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,65536,0.509552001953125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,16384,0.3866124418046739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,12288,0.07246222098668416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,10240,0.023171555664804246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,12288,0.018242667118708294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,12288,0.33144622378879124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,8192,0.01807822287082672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,10240,0.01623911162217458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,10240,0.36945332421196836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,8192,0.04820533262358772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,7168,0.01645511057641771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,8192,0.013762666947311826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,8192,0.35001333554585773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,7168,0.043416887521743774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,6144,0.015076445208655464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,7168,0.012788444757461548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,7168,0.3709866735670302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,6144,0.03872088922394647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,6144,0.344920900132921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,5120,0.013285333083735572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,6144,0.011493333511882357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,10240,0.060538669427235924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,5120,0.03502044412824843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,5120,0.01035377797153261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,4096,0.011847111086050669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,5120,0.3519022199842665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,4096,0.028840889533360798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,3584,0.010089777410030365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,4096,0.009370666411187913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,4096,0.3607066737280951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,3584,0.026434666580624048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,3072,0.009360888765917884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,3584,0.008773333496517604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,3584,0.35005688667297363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,3072,0.02453777856296963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,3072,0.008365333080291748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,3072,0.3149777783287896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,2560,0.021352888809310064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,2560,0.33182843526204425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,2048,0.007736889024575551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,2048,0.018918222851223417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,2560,0.008291555775536431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,2048,0.3798275523715549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,1536,0.006343110981914733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,2048,0.0074675555030504865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,1536,0.01721244388156467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,2560,0.007975111405054728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,1024,0.005231110999981563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,1536,0.0069413334131240845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,1536,0.34892800119188094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,1024,0.01480888823668162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,768,0.004919111314747068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,1024,0.3378017743428548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,768,0.014087999860445658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,512,0.004576888763242298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,768,0.006311999840868845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,512,0.013478221992651621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,768,0.3283991018931071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,256,0.0038844446341196695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,512,0.006280000011126201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,256,0.013085333009560903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,512,0.37279465463426376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,128,0.0038355555799272326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,256,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,256,0.3292320039537218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,1024,0.006578666468461354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,128,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2560,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,32,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2560,128,0.3079306549496121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,64,0.01275111072593265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,65536,0.09519822067684597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2560,32,0.012437333663304647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,65536,0.07356977462768555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,65536,0.325163549847073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,16384,0.027035555905765955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,65536,0.4944799741109212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,16384,0.08271021975411309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,12288,0.023181334137916565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,16384,0.3330204486846924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2560,64,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,16384,0.02202488813135359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,10240,0.020748444729381137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,12288,0.018047110901938546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,12288,0.06448266903559367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,12288,0.3195173210567898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,10240,0.05362488826115926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,8192,0.01588000026014116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,10240,0.015619556109110514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,10240,0.3240719901190864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,8192,0.04433777928352356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,7168,0.014953777194023132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,8192,0.31770311461554634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,7168,0.03975111246109009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,6144,0.013748444616794586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,7168,0.012430222498046027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,7168,0.32983022265964085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,6144,0.03609600000911289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,5120,0.012063110868136087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,6144,0.011389333340856763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,6144,0.3121866650051541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,5120,0.03193599979082743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,4096,0.010407999985747868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,5120,0.010194666683673859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,5120,0.2985404332478841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,4096,0.026879999372694228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,4096,0.009302222066455418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,4096,0.31874222225613064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,3584,0.024311110377311707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,3584,0.00869955536392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,3584,0.33309955067104763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,3072,0.008591111335489485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,3072,0.02232444451914893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,3072,0.3326764371660021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,3072,0.00833155545923445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,8192,0.013172444370057849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,2560,0.35886398951212567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,2560,0.007305777735180325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,2560,0.020420443680551317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,2560,0.008014221986134848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,2048,0.006677333265542984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,2048,0.01758311192194621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,2048,0.363371557659573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,2048,0.007283555964628856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,1536,0.005571555760171678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,1536,0.01588088936275906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,1536,0.34765156110127765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,1536,0.007009777757856581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,1024,0.004915555732117759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,3584,0.009708444277445475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,1024,0.014460444450378418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,768,0.004599110947714912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,1024,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,1024,0.32707288530137807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,768,0.013822221921549903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,512,0.004224888980388641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,768,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,768,0.343722661336263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,512,0.013343111508422427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,256,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,512,0.3280257648891873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,256,0.012786666552225748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,256,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,128,0.0038853333228164246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,128,0.01257333325015174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,128,0.30524444580078125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,64,0.0038622220357259116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,128,0.006020444548792309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,2048,32,0.0036657779581016968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,64,0.012500444220172035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,65536,0.07910222477383085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,2048,32,0.012215110990736218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,2048,256,0.34276888105604386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,65536,0.0722293323940701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,65536,0.314667542775472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,16384,0.027278222971492346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,65536,0.4950186411539714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,16384,0.08192355765236749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,12288,0.019127999742825825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,16384,0.3316631052229139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,2048,512,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,16384,0.021770666042963665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,10240,0.01699022286468082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,12288,0.06263022290335761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,12288,0.01723022262255351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,12288,0.31500087844000924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,10240,0.015117333994971381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,10240,0.051763554414113365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,10240,0.33225777414109975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,8192,0.04280622137917412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,7168,0.012681778106424542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,8192,0.01254488858911726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,8192,0.39752356211344403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,7168,0.038689778910742864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,6144,0.011736888852384357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,7168,0.29619288444519043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,7168,0.011797332929240333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,6144,0.03490666548411051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,5120,0.010195555786291758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,6144,0.011056000159846412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,6144,0.30456799930996364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,8192,0.01406133340464698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,5120,0.03057866626315647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,4096,0.009659555223253038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,5120,0.010379555324713389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,5120,0.3820684485965305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,4096,0.025382222400771245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,4096,0.009549332989586724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,3584,0.008755555583371056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,4096,0.31132533815171987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,3584,0.023500444160567388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,3584,0.008738666772842407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,3584,0.3134640057881673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,3072,0.021731555461883545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,2560,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,3072,0.008351999852392409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,3072,0.37105245060390896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,2560,0.019916445016860962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,2048,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,2560,0.00796622203456031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,2560,0.31522488594055176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,2048,0.01752622259987725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,1536,0.005608888963858287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,2048,0.00758844448460473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,1536,0.016107555892732408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,2048,0.3710373242696126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,3072,0.007768889268239339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,1024,0.00479911102188958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,1536,0.3180977768368191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,1024,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,768,0.004297777596447202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,1024,0.00658044425977601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,1024,0.36678044001261395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,768,0.01344622257683012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,512,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,768,0.006301333506902059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,768,0.3174933327568902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,512,0.013075555364290873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,512,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,256,0.003559111307064692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,512,0.3628097905053033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,1536,0.006958222223652734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,256,0.01260444439119763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,128,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,256,0.35399733649359805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,256,0.005896889087226655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,128,0.01237511138121287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,64,0.003546666767862108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1536,128,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1536,32,0.00318755561278926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,64,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1536,128,0.34288978576660156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,65536,0.05240177777078417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1536,32,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,65536,0.07038488652971056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,65536,0.2751839955647786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,16384,0.01906044450071123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,65536,0.4905742009480794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,16384,0.07120888763003878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,12288,0.018539556198649935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,16384,0.3643173376719157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,16384,0.020282665888468426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,12288,0.053358223703172475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,12288,0.016005333926942613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,10240,0.01564800077014499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,12288,0.3466177781422933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,10240,0.044960889551374644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,8192,0.013272888958454132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,10240,0.015184889237085978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,10240,0.3497768772972955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,8192,0.037666665183173284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,7168,0.012280888855457306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,8192,0.012473777764373355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,8192,0.307257784737481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,7168,0.0336426662074195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,7168,0.011753777662913004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,6144,0.010785777535703448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,7168,0.3666168848673503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,6144,0.031140443351533677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,5120,0.009510222408506606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,6144,0.343739562564426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,5120,0.02752799954679277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,5120,0.33362399207221133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,4096,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,4096,0.0229759994480345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,4096,0.3329288959503174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,4096,0.009240888886981541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,6144,0.011001778145631155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,3584,0.007461333440409766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,3584,0.021354667014545862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,3584,0.00865955568022198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,3072,0.006609777609507243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,5120,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,3584,0.3682542112138536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,3072,0.01993600030740102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,2560,0.005984888722499211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,3072,0.008296888735559251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,3072,0.31242044766743976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,2048,0.005511111269394557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,2560,0.007683555285135905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,2048,0.016551110479566786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,2560,0.3576177755991618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,1536,0.005061333378156026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,2048,0.007357333269384172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,2048,0.32309865951538086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,1536,0.015141333142916361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,1024,0.004574222283230888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,1536,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,1536,0.32766665352715385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,1024,0.01347644461525811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,768,0.003945777813593547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,1024,0.00666844430896971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,1024,0.3031244542863634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,2560,0.01832444469134013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,768,0.01315111087428199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,512,0.003899555653333664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,768,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,768,0.3233368926578098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,256,0.0032311110860771606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,512,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,512,0.3162657684750027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,256,0.012175111307038201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,128,0.0032071111102898917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,256,0.005906666732496685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,256,0.38013066185845273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,128,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,64,0.0032755554550223877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,128,0.005943111247486538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,1024,128,0.29872798919677734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,1024,32,0.0032097777972618737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,64,0.011744889120260874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,65536,0.04811822374661764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,1024,32,0.011756444142924415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,65536,0.2739848825666639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,65536,0.06768266359965007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,16384,0.017514665921529133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,65536,0.48670313093397355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,16384,0.06979999939600627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,16384,0.3025884363386366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,16384,0.01904000010755327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,12288,0.015360888507631091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,12288,0.05241866575347053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,12288,0.33371734619140625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,10240,0.014096889230940076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,12288,0.015822221835454304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,1024,512,0.005922666854328579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,10240,0.045019557078679405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,8192,0.013051555388503604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,10240,0.014434667097197639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,10240,0.3459004561106364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,8192,0.036751998795403376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,7168,0.011786667009194693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,8192,0.3423413435618083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,7168,0.03297422329584757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,7168,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,7168,0.3466373284657796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,6144,0.010524444282054901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,6144,0.0302746660179562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,6144,0.3503137694464789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,6144,0.011057777537239922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,5120,0.009022222624884712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,8192,0.012392000191741519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,5120,0.026829333768950567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,4096,0.00812000036239624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,5120,0.010043555663691627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,5120,0.3511502212948269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,4096,0.023035556077957153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,4096,0.009002666506502364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,3584,0.007356444166766272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,4096,0.336151123046875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,3584,0.021233777205149334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,3072,0.0069395556218094295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,3584,0.008722667064931657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,3584,0.3378133243984646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,3072,0.019907555646366544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,2560,0.00630044440428416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,3072,0.008323555191357931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,3072,0.33072710037231445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,2048,0.005589333259397083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,2560,0.018593778212865193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,2560,0.31449956364101833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,2048,0.015844444433848064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,2560,0.008040000167157915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,1536,0.005120888766315249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,2048,0.007272000114123027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,1536,0.01494311127397749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,1536,0.006937777830494775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,1536,0.32510844866434735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,1024,0.004639111045334074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,1024,0.013510222236315409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,768,0.004231111043029361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,1024,0.006693333387374878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,1024,0.31839556164211696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,768,0.013086222112178802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,512,0.0038684445122877755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,768,0.006257777826653586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,512,0.012757333616415659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,768,0.4438329007890489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,256,0.0035253332720862497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,2048,0.32987909846835667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,512,0.3191280100080702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,512,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,128,0.0032248888164758682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,256,0.0058853332367208265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,256,0.012164444559150271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,128,0.012134222520722283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,256,0.32595645056830513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,64,0.003170666595300039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,768,32,0.0033413333197434745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,768,128,0.005930666708283954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,65536,0.03396888905101352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,768,128,0.2944515546162923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,64,0.0118222220076455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,768,32,0.012102222277058495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,16384,0.014490667316648694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,65536,0.06753955284754436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,65536,0.2780560122595893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,65536,0.4890853563944499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,12288,0.013613333304723104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,16384,0.3293173313140869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,16384,0.01897333396805657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,12288,0.05241688754823473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,12288,0.015950222810109455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,12288,0.3272293408711751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,10240,0.04371111260520088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,10240,0.3252248764038086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,10240,0.014226666755146451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,8192,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,16384,0.06775022215313382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,8192,0.03602666656176249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,7168,0.011086222198274402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,8192,0.012447111308574677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,7168,0.0324942237801022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,10240,0.012424888710180918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,6144,0.010038221875826517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,7168,0.3329039944542779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,7168,0.011728000309732227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,6144,0.02921600143114726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,6144,0.3383617666032579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,6144,0.01110400011142095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,5120,0.009018667042255402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,5120,0.0262746661901474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,8192,0.326690673828125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,5120,0.34618756506178117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,4096,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,5120,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,4096,0.023028444912698533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,3584,0.007284444239404466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,4096,0.35756267441643613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,3584,0.021393777595625982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,3584,0.008665777742862701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,3072,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,3072,0.019911999503771465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,4096,0.009004444711738164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,3072,0.3510897689395481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,2560,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,2560,0.018274666534529794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,3584,0.34612178802490234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,2560,0.007735110819339752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,2048,0.0052639999323421055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,2560,0.3540168868170844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,2048,0.016208888755904306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,2048,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,2048,0.3423982196384006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,1536,0.004893333133724001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,3072,0.008332444561852349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,1536,0.014883556299739413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,1536,0.006966222491529252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,1024,0.004287111262480418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,1536,0.31632177035013836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,1024,0.013439111411571503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,768,0.0039022221333450745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,1024,0.006735111276308696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,1024,0.32663199636671275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,768,0.01309511148267322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,512,0.0035111109415690103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,768,0.006372444331645966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,512,0.012448888685968188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,256,0.0035093331502543557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,512,0.006055999961164262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,256,0.012056888805495368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,256,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,128,0.002942222274012036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,256,0.42905333307054305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,128,0.012087110843923358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,512,128,0.005984000033802456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,768,0.3148062229156494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,128,0.37308798895941836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,512,32,0.002891555635465516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,65536,0.0244186669588089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,64,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,512,512,0.3186693456437853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,512,32,0.0118231111102634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,16384,0.010420444111029306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,65536,0.06619733572006226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,65536,0.2764417860243055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,16384,0.06811288992563884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,65536,0.4924355612860785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,12288,0.008752000000741746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,16384,0.018917333748605516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,16384,0.32707201109992134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,12288,0.05216799841986763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,12288,0.015620445211728414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,10240,0.011078221930397881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,10240,0.014135999812020196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,12288,0.3854542308383518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,10240,0.04278666774431864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,8192,0.009757333331637913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,10240,0.32743112246195477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,7168,0.009378666679064432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,8192,0.0124177775449223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,8192,0.03615022367901272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,8192,0.3251270982954237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,6144,0.009003555609120263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,7168,0.011818666425016193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,7168,0.31869954533047146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,6144,0.02920711040496826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,5120,0.008283555507659912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,6144,0.0107351111041175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,6144,0.36511821217007107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,5120,0.026504889130592346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,4096,0.007296000089910295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,5120,0.010241777532630498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,5120,0.32390933566623265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,4096,0.022671999202834234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,3584,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,4096,0.008995555341243744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,3584,0.021273778544531927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,4096,0.3656488789452447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,7168,0.032491554816563926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,3072,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,3584,0.008638222184446123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,3584,0.32286132706536186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,2560,0.005912000106440649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,3072,0.008314666648705801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,3072,0.35660354296366376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,2560,0.017901332841979135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,2560,0.0076453329788313965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,2048,0.0052577778697013855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,2560,0.3596497906578912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,2048,0.015832000308566622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,2048,0.35736444261338973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,2048,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,1536,0.004919111314747068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,3072,0.01961599952644772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,1536,0.014472888575659858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,1024,0.003919110943873723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,1536,0.007010666860474481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,1536,0.3339244524637858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,1024,0.013105777402718862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,768,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,1024,0.006331555545330048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,1024,0.3494577672746446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,768,0.012424888710180918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,512,0.0035217776894569397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,768,0.33915466732449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,512,0.012162666353914471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,512,0.006015111174848344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,512,0.3404835595024957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,256,0.003216000066863166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,256,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,256,0.3603999879625108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,256,0.005917333480384614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,128,0.0031999999450312722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,128,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,768,0.006246222390068903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,64,0.0029253332565228143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,256,128,0.3120737870534261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,256,128,0.005930666708283954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,256,32,0.0028284444577164124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,65536,0.015484443969196744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,64,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,256,32,0.011776000261306763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,65536,0.0646044413248698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,65536,0.27686844931708443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,16384,0.00998488896422916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,65536,0.48882044686211484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,16384,0.01903377804491255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,16384,0.06740266746944852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,12288,0.05024266574117872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,16384,0.5156133439805772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,12288,0.01581333412064446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,10240,0.008700444466537898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,10240,0.04222933451334635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,10240,0.014168888330459595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,8192,0.007987555530336168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,10240,0.4620809025234646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,12288,0.008357333640257517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,8192,0.03536622060669793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,7168,0.007621333003044128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,8192,0.012453333371215396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,8192,0.4201768769158258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,7168,0.03228800164328681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,12288,0.3251875506507026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,6144,0.007368889119890001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,7168,0.32589599821302623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,7168,0.011847111086050669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,6144,0.029511110650168523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,5120,0.007752888732486301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,6144,0.010801778071456485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,6144,0.3246871100531684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,5120,0.0258559998538759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,4096,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,5120,0.010067555639478896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,4096,0.022322666313913133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,5120,0.38483378622266984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,4096,0.00905777762333552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,4096,0.3181946542527941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,3584,0.021016889148288306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,3584,0.008700444466537898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,3584,0.3063431051042345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,3072,0.019660444723235238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,3072,0.31329067548116046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,3072,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,2560,0.00563822231358952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,2560,0.018217777212460835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,2560,0.3089395629035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,2048,0.0052826665341854095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,2048,0.01593955523437924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,3072,0.006603555546866522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,2048,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,1536,0.00462755560874939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,2048,0.3287759886847602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,1536,0.01482399966981676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,1536,0.3262968858083089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,1024,0.0038622220357259116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,1536,0.007032000356250339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,3584,0.0069368887278768755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,2560,0.007723555796676212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,768,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,1024,0.013201778133710226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,1024,0.006324444380071428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,768,0.012768000364303589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,1024,0.36366934246487087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,512,0.00319377767542998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,768,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,768,0.32687200440300834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,512,0.012352888782819113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,256,0.0032666667054096856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,512,0.0059573331640826324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,512,0.34859466552734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,256,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,256,0.0059368887709246735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,128,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,256,0.350524451997545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,128,0.011734222372372946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,64,0.0028604444944196274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,512,128,128,0.005973333285914526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,128,32,0.0029351111087534162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,64,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,65536,0.0148053334818946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,128,32,0.011406222151385413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,16384,0.009381333159075843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,16384,0.06650488906436496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,65536,0.27741776572333443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,12288,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,10240,0.00904622260067198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,12288,0.049992889165878296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,8192,0.0083155557513237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,10240,0.04179822074042427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,7168,0.00794666674402025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,8192,0.03535555468665229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,6144,0.007311999797821045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,7168,0.032256888018714056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,5120,0.007684444387753804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,6144,0.02917333443959554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,4096,0.007289778027269576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,4096,0.022675555613305833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,3584,0.0070257774657673305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,3584,0.02124622298611535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,3072,0.006960888703664144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,3072,0.019681778219011094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,2560,0.006559999866618051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,5120,0.02610844373703003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,2048,0.006234666539563074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,2560,0.01827822294500139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,1536,0.004538666870858935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,2048,0.015990222493807476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,1024,0.0038924444880750445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,1536,0.01479111115137736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,768,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,1024,0.013131555583741931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,512,0.0032337777730491427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,768,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,256,0.0031884445084465873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,128,0.002938666691382726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,256,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,128,0.011521778172916837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,64,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,32,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,65536,0.0185217774576611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,65536,0.2773679892222087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,16384,0.009000000026490953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,16384,0.06723822487725152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,12288,0.008360000120268928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,64,32,0.002595555658141772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,10240,0.007734222544564142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,12288,0.04978311061859131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,8192,0.00796622203456031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,10240,0.04150933358404372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,7168,0.008053333395057255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,8192,0.03535999854405721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,6144,0.007371555599901412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,7168,0.031934221585591636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,64,512,0.012085333466529846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,5120,0.007304000357786815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,6144,0.029123554627100628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,4096,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,3584,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,5120,0.026162667406929865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,4096,0.02274755636850993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,3072,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,2560,0.006637333167923822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,3584,0.02070755594306522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,2048,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,3072,0.019658666517999437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,1536,0.0045422220395671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,2560,0.01822933389080895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,1024,0.003908444609906939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,2048,0.015849777393870883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,768,0.003569777641031477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,1536,0.014868444866604276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,512,0.0032053333189752367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,768,0.012451555165979596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,256,0.003265777809752358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,512,0.01240533341964086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,128,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,256,0.011686222420798408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,64,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,128,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,512,32,32,0.00281866660548581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,64,0.011426666544543372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,32,0.011726222104496427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,512,32,1024,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,16384,0.4344488779703776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,16384,0.5248577859666612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,16384,0.20997867319318983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,12288,0.3414319886101617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,12288,0.4280213250054254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,16384,0.6297600004408095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,12288,0.15578222274780273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,10240,0.2904124524858263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,12288,0.4956551127963596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,10240,0.40680623054504395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,10240,0.13201244672139487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,8192,0.22772267129686144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,10240,0.428438213136461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,8192,0.2679111162821452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,7168,0.21525333987341988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,7168,0.2356088956197103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,8192,0.4014773368835449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,7168,0.09676888916227554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,8192,0.10853777991400824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,6144,0.1876951058705648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,7168,0.3470773431989882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,6144,0.2000337839126587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,6144,0.08445955647362603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,5120,0.14684177769554987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,6144,0.29121777746412486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,5120,0.07398311297098796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,4096,0.13048088550567627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,4096,0.14444266425238714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,5120,0.17368355062272814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,4096,0.0602782236205207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,4096,0.3252791033850776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,3584,0.10808000299665664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,3584,0.1276622215906779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,3584,0.05422311027844747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,3584,0.3112728860643175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,3072,0.09741689099205865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,3072,0.10744888914955987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,5120,0.3056737846798367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,2560,0.0803333322207133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,2560,0.09233511156506008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,2560,0.3355591032240126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,3072,0.34338755077785915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,3072,0.04783911175198025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,2048,0.06715022193060981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,2048,0.08008178075154622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,2048,0.03696177734269036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,2048,0.32658667034573025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,1536,0.048355556196636625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,1536,0.06505155563354492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,1536,0.33470932642618817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,2560,0.04285155402289497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,1024,0.035828444692823626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,1024,0.05017155408859253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,1024,0.32085243860880536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,768,0.029845333761639063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,1024,0.02254399988386366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,768,0.04308977723121643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,768,0.020053333706325956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,768,0.33000532786051434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,512,0.023400889502631292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,1536,0.030282666285832722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,512,0.03679200013478597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,256,0.01533777846230401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,512,0.018353778454992507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,512,0.32897154490152997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,256,0.03221244282192654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,128,0.013584000368913015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,256,0.32701155874464244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,128,0.030571556753582422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,65536,128,0.29378043280707467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,128,0.01684888866212633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,64,0.01520088811715444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,65536,32,0.015820443630218506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,64,0.029903110530641343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,65536,32,0.03218044506178962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,65536,256,0.017118222183651395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,65536,0.46091201570298934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,65536,0.6681075625949435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,16384,0.12793510489993626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,65536,0.23902667893303764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,16384,0.17484445042080352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,16384,0.06582488616307576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,16384,0.3457982275221083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,12288,0.09688444270028009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,12288,0.13367999924553767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,12288,0.049709333313835986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,12288,0.3383653428819444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,10240,0.09502310885323419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,65536,0.8795662456088595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,10240,0.11205332809024388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,8192,0.07954666349622938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,10240,0.04351377818319532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,10240,0.331862211227417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,8192,0.09257422553168403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,7168,0.07136888636483087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,8192,0.33351288901435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,7168,0.08306755622227986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,7168,0.032487110959159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,7168,0.3108666737874349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,6144,0.0624835557407803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,6144,0.0718924469417996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,6144,0.34906578063964844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,8192,0.036333332459131874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,5120,0.05163822240299649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,6144,0.029177778297000464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,5120,0.061252441671159535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,5120,0.3396577835083008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,4096,0.043140444490644664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,4096,0.05180977781613668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,4096,0.3369893232981364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,512,128,128,0.3395715554555257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,4096,0.020925333102544148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,3584,0.037104000647862755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,5120,0.02427111069361369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,3584,0.048179556926091514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,3584,0.019105777144432068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,3072,0.03351733419630263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,3584,0.34928001297844785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,3072,0.04155200057559543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,2560,0.028755555550257366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,3072,0.017009778155220878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,3072,0.31736622916327584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,2560,0.03521777855025397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,2560,0.015992888145976596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,2048,0.024192000428835552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,2560,0.3418284522162543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,2048,0.030951112508773804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,1536,0.018295110927687753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,2048,0.014133334159851074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,2048,0.3240142133500841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,1536,0.026424888107511733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,1536,0.01184177812602785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,1024,0.013431111143694984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,1536,0.3356906572977702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,1024,0.021951999929216173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,768,0.011044444309340583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,1024,0.010030222435792288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,1024,0.3352515432569716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,768,0.019716444942686293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,512,0.008752889103359645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,768,0.009317333499590555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,768,0.3236568768819173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,512,0.01759822169939677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,256,0.006583999842405319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,512,0.00869066682126787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,512,0.33108266194661456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,256,0.016606221596399944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,128,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,256,0.008364444805516137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,256,0.32836887571546763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,64,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,128,0.01626133256488376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,16384,32,0.006664000037643645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,16384,128,0.30350756645202637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,64,0.016188444362746347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,16384,32,0.016512889001104567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,65536,0.34725332260131836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,65536,0.5417013168334961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,16384,0.09485689136717056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,65536,0.1890559991200765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,16384,0.14300533135732016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,16384,128,0.008571555217107138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,12288,0.07161600059933133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,16384,0.05193066596984863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,16384,0.3268960052066379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,12288,0.10921421978208755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,10240,0.06905066967010498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,12288,0.04063911239306132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,65536,0.7114666832817925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,12288,0.3414035638173421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,10240,0.09145689010620117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,8192,0.057773331801096596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,10240,0.034494221210479736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,10240,0.32518310017055935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,8192,0.07542310820685492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,8192,0.028615110450320776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,7168,0.050740443997912936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,8192,0.3212195502387153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,6144,0.04380711250834995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,7168,0.0674826635254754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,7168,0.32794398731655544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,6144,0.06003466579649183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,5120,0.038359999656677246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,6144,0.3070017761654324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,6144,0.022688888841205176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,5120,0.05056711037953695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,5120,0.019871999820073444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,4096,0.03164800008138021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,5120,0.3119093312157525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,4096,0.042895111772749156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,3584,0.028743998871909246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,4096,0.017447110679414537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,4096,0.3459324571821425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,7168,0.025889777474933203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,3584,0.03794044587347243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,3072,0.02446844511561924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,3584,0.015923556354310777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,3584,0.32659022013346356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,3072,0.033377776543299355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,2560,0.02100444502300686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,3072,0.014596444037225513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,3072,0.3255946636199951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,2560,0.03012622065014309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,2048,0.017476444443066914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,2560,0.3337457709842258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,2560,0.013526221944226159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,2048,0.02607022225856781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,1536,0.013838222457302941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,2048,0.011783999701340994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,2048,0.32580089569091797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,1536,0.02310133311483595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,1024,0.010954666468832227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,1536,0.010062221851613786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,1536,0.341031101014879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,1024,0.019526221685939364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,768,0.009156444834338294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,1024,0.009082666701740688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,1024,0.3409840001000299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,768,0.017715555098321702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,512,0.007291555404663086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,768,0.33841511938307023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,768,0.008301333420806462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,512,0.016184000505341422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,256,0.005615111026499007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,512,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,512,0.3374977906545003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,256,0.015494222442309061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,128,0.00489155575633049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,256,0.007616000043021307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,256,0.3551173210144043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,128,0.015138667490747241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,64,0.004565333326657613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,12288,128,0.007631111476156447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,12288,128,0.3073964383867052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,12288,32,0.004852444347408083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,64,0.015083556373914083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,12288,32,0.01574577722284529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,65536,0.2784595489501953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,65536,0.4908160103691949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,16384,0.07699288924535115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,65536,0.18273954921298555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,16384,0.12844710879855686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,16384,0.04934844374656677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,16384,0.34301598866780597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,12288,0.06140622165468004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,12288,0.10092000166575114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,65536,0.7175679736667209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,10240,0.052332444323433765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,12288,0.03884177737765842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,12288,0.3320755428738064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,8192,0.04333244429694282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,10240,0.03343288766013251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,10240,0.08505600028567845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,10240,0.34159644444783527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,8192,0.06951822174919976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,7168,0.03830577929814657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,8192,0.32773510615030926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,7168,0.062255111005571157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,6144,0.032947556840048894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,7168,0.024455111887719896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,7168,0.3343591160244412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,6144,0.054542223612467446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,6144,0.02204977803760105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,5120,0.029616001579496596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,6144,0.32612087991502553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,5120,0.019447111421161227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,5120,0.046562665038638644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,4096,0.025156444973415796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,5120,0.32096976704067653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,8192,0.027775999572541978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,3584,0.022633777724372015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,4096,0.03956088754865859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,4096,0.016726222303178575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,4096,0.3079555564456516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,3072,0.019690665933820937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,3584,0.0155102229780621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,3584,0.3297911220126682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,3072,0.03145866592725118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,2560,0.01735466718673706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,3072,0.3275093237559001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,3072,0.014115555418862237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,2560,0.028157333532969158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,2560,0.012888000243239932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,2048,0.014922666880819531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,2560,0.32436267534891766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,2048,0.02472977836926778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,1536,0.012158222496509552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,2048,0.01110311100880305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,2048,0.3896924389733209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,3584,0.036537776390711464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,1536,0.02167111138502757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,1024,0.010152888794740042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,1536,0.3177866670820448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,1536,0.009735110733244155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,768,0.008468444148699442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,1024,0.008647999829716152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,1024,0.32650666766696507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,768,0.01687111126052009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,512,0.006924444602595435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,768,0.008270222279760573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,768,0.3299822277492947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,512,0.015557333827018738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,512,0.007990222010347579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,512,0.3394560019175212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,256,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,256,0.015136889285511441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,1024,0.018657777044508193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,256,0.3572239875793457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,256,0.007616000043021307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,128,0.005251555393139522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,10240,128,0.007336888876226213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,64,0.004915555732117759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,10240,128,0.3089262114630805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,10240,32,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,64,0.014429334137174817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,32,0.015290665957662793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,65536,0.28306489520602757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,65536,0.44159555435180664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,10240,128,0.014505777094099255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,16384,0.0787608888414171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,65536,0.13433955775366888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,16384,0.11597600248124863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,16384,0.03794222076733907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,65536,0.5799022250705296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,16384,0.33090045717027455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,12288,0.06155911419126722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,12288,0.02991022335158454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,10240,0.051474667257732816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,12288,0.33677156766255695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,10240,0.07660444577534993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,10240,0.3402426772647434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,10240,0.025377778543366328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,8192,0.04279822111129761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,8192,0.06335022052129109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,8192,0.3328915437062581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,12288,0.09092266692055596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,7168,0.0380639996793535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,7168,0.05664088990953234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,6144,0.03322044346067641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,7168,0.32126132647196454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,7168,0.019766221443812054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,6144,0.049363556835386485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,5120,0.028528001573350694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,6144,0.33235645294189453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,6144,0.017637333936161466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,5120,0.04203022188610501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,4096,0.024247111545668706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,5120,0.01570311188697815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,8192,0.02145599987771776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,5120,0.3229244550069173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,3584,0.021344888541433547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,4096,0.036419557200537786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,4096,0.013795555465751223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,4096,0.3136533366309272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,3584,0.03350755572319031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,3072,0.017630222770902846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,3584,0.012788444757461548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,3584,0.32404179043240017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,2560,0.015769778026474845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,3072,0.029198222690158423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,3072,0.011763555308183035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,3072,0.31347645653618705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,2560,0.026131555438041687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,2560,0.010653333531485664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,2560,0.33823111322191024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,2048,0.022981334063741896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,1536,0.011559999651379056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,2048,0.00963199966483646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,2048,0.3362560007307265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,1536,0.02033511135313246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,1024,0.009268444445398118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,1536,0.008735111190213097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,1536,0.34014222357008195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,1024,0.017179555363125272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,768,0.00777599960565567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,1024,0.007948444121413762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,1024,0.3226657708485921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,768,0.016173332929611206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,512,0.006321777900060018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,2048,0.013756444056828817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,768,0.007677333222495184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,512,0.014807111687130399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,768,0.33339288499620223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,512,0.007084444165229797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,512,0.2992773320939806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,256,0.014482667048772177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,128,0.004917333523432414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,256,0.006597333484225803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,256,0.33074845208062065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,128,0.014257777068350049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,64,0.00453599997692638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,8192,128,0.006649777707126405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,32,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,8192,128,0.2939679887559679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,64,0.013808888693650564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,8192,32,0.014156444205178155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,65536,0.2730426788330078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,8192,256,0.005254222287072076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,16384,0.0762702226638794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,65536,0.40251021915011936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,65536,0.13055377536349827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,16384,0.10791022247738308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,16384,0.29858043458726674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,16384,0.037273777855767146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,12288,0.05888355440563626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,65536,0.5714826583862305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,12288,0.08296533425649007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,10240,0.04270844327078926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,12288,0.32802311579386395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,12288,0.028579556279712256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,10240,0.06975999805662367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,10240,0.02490844494766659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,8192,0.03511022196875678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,10240,0.3231733375125461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,8192,0.05828711059358385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,7168,0.032264888286590576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,8192,0.02110044492615594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,8192,0.31848711437649196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,7168,0.05139466789033678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,7168,0.019544000426928203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,6144,0.028484443823496502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,7168,0.3312053415510389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,6144,0.04499466551674736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,6144,0.017350221673647564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,6144,0.3484800126817491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,5120,0.039134220944510564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,5120,0.015524443652894763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,5120,0.32725244098239475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,4096,0.020486222373114694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,4096,0.03408711155255636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,4096,0.34286043379041886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,4096,0.013831111292044321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,3584,0.018053332964579265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,5120,0.023950222465727065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,3584,0.03083555565940009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,3072,0.01566755606068505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,3584,0.012427555190192329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,3584,0.3417670991685655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,3072,0.027152000202072993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,2560,0.01384622189733717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,3072,0.011432888607184092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,3072,0.3385475476582845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,2560,0.024561777710914612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,2048,0.012160888976520963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,2560,0.0105262224872907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,2560,0.33073422643873424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,2048,0.021996445126003687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,2048,0.00942666663063897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,2048,0.32880621486239964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,1536,0.019222221440739103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,1024,0.00775022225247489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,1536,0.008653333617581261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,1536,0.3373102347056071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,1024,0.01640533407529195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,768,0.006710222197903528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,1024,0.007975111405054728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,1024,0.3367946677737766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,768,0.015469332536061605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,512,0.005578666511509154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,768,0.007307555940416124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,768,0.32265154520670575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,512,0.014522666732470194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,256,0.004912000149488449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,512,0.006949333681000604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,512,0.32682932747734916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,256,0.014111111561457316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,128,0.0042444442709287005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,256,0.006597333484225803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,128,0.013728888498412238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,7168,128,0.006694222076071634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,128,0.29401156637403697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,64,0.003927999900446998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,32,0.004256000121434529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,64,0.01346666696998808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,7168,32,0.013819555441538492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,7168,1536,0.009859555297427708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,65536,0.2230471240149604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,65536,0.36633510059780544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,65536,0.12746666537390813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,16384,0.09700977802276611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,65536,0.4379448890686035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,16384,0.03529511226548089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,16384,0.338010655509101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,12288,0.046739555067486234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,12288,0.0748808913760715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,7168,256,0.3337102201249864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,10240,0.040957334968778826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,16384,0.0608160032166375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,12288,0.2991733286115858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,12288,0.027469333675172593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,8192,0.03429955575201247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,10240,0.06377155250973172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,10240,0.023863111933072407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,10240,0.3129884401957194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,8192,0.05225422316127353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,7168,0.02933599882655674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,8192,0.02052888936466641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,8192,0.3127680089738634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,7168,0.04684977730115255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,7168,0.01854044364558326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,7168,0.3179653220706516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,6144,0.04115110966894362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,6144,0.3034115632375081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,6144,0.016935999194780987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,5120,0.03534222311443753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,5120,0.34900445408291286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,6144,0.02676533328162299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,4096,0.018479110466109384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,5120,0.015196444259749519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,4096,0.03130844566557143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,5120,0.0218231115076277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,4096,0.013285333083735572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,3584,0.015457777513398064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,4096,0.34699198934766984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,3584,0.028824888997607764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,3584,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,3072,0.01482222146458096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,3584,0.3073786629570855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,3072,0.02535466684235467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,2560,0.012696888711717395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,3072,0.011026666396194033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,3072,0.33450667063395184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,2560,0.022823111878501043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,2560,0.01034311122364468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,2048,0.011227555572986603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,2560,0.31358933448791504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,2048,0.020610666937298246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,1536,0.009157333109113906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,2048,0.008982222113344405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,2048,0.32751022444831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,1536,0.017838221457269456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,1024,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,1536,0.2980959945254856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,1024,0.015822221835454304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,1024,0.29721421665615505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,1024,0.007692444655630324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,768,0.006278222219811545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,768,0.01519466605451372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,768,0.30316000514560276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,512,0.004914666629499859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,512,0.013857777747843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,1536,0.00831200016869439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,256,0.00423200014564726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,512,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,256,0.013791999883121915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,512,0.48448530832926434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,128,0.0038986665507157645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,256,0.006576889091067844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,256,0.31095200114780003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,128,0.013479111095269522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,64,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,6144,128,0.30113066567314994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,768,0.007295110987292395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,6144,32,0.003906666818592283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,64,0.01313244468635983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,6144,32,0.013633777697881063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,65536,0.19030133883158365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,65536,0.3503031200832791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,65536,0.1260862218009101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,16384,0.051939556996027626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,16384,0.09493777487013076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,6144,128,0.006620444357395172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,65536,0.43847910563151044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,16384,0.33408533202277285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,16384,0.033477332856920027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,12288,0.07412355475955539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,12288,0.026203556193245783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,12288,0.32152001063028973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,10240,0.06144533554712931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,10240,0.32525955306159127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,10240,0.022919111781650122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,8192,0.029522667328516643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,8192,0.050738665792677135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,12288,0.039339555634392634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,7168,0.025027554896142747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,8192,0.33764976925320095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,8192,0.019560888409614563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,7168,0.044176889790429004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,10240,0.03580800029966567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,6144,0.022287999590237934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,7168,0.32099999321831596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,6144,0.039770666095945574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,6144,0.016363556186358135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,5120,0.019490665859646268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,6144,0.31625244352552623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,5120,0.03464533223046197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,4096,0.016163556112183463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,5120,0.014663111832406787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,5120,0.31748533248901367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,4096,0.0303439994653066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,4096,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,4096,0.32240978876749676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,7168,0.01796444422668881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,3584,0.028146667612923518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,3072,0.01349244432316886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,3584,0.011773332953453064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,3584,0.2897111044989692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,3072,0.025038222471872967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,2560,0.011772444678677453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,3072,0.010692444112565784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,3072,0.31276710828145343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,2560,0.022355554832352534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,2560,0.31260622872246635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,2560,0.009730666875839233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,2048,0.01001422190003925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,3584,0.014679999815093147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,2048,0.01963644391960568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,1536,0.008596444295512306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,2048,0.00869777798652649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,2048,0.30792533026801217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,1536,0.01733688844574822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,1024,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,1536,0.33184978697035045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,1024,0.01576622161600325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,1024,0.3361635473039415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,1024,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,768,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,768,0.014880888991885714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,768,0.33972978591918945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,768,0.007350222104125553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,512,0.00507733349998792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,1536,0.007982222570313348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,512,0.014112000664075216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,256,0.004558222161398994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,512,0.324743111928304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,256,0.013426666458447775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,256,0.006583111153708563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,256,0.3265484439002143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,128,0.013078221844302284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,5120,128,0.32187289661831325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,512,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,5120,128,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,64,0.003863111138343811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,32,0.0038204445607132385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,64,0.012834666503800286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,5120,32,0.013038222160604266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,5120,128,0.0038657776183552216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,65536,0.15003466606140137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,65536,0.3119991090562609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,65536,0.07584533426496717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,16384,0.040885332557890154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,16384,0.08426933156119452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,16384,0.3342657883961995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,16384,0.02254399988386366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,12288,0.03241689006487528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,12288,0.06500177913241915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,10240,0.02825688984658983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,12288,0.01833422150876787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,10240,0.054069333606296115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,10240,0.3114275667402479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,10240,0.01648266613483429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,8192,0.022919111781650122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,65536,0.43426044782002765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,8192,0.04439822170469496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,7168,0.0203146669599745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,8192,0.014436443646748861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,8192,0.3168533378177219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,7168,0.040143112341562905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,7168,0.012886222038004132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,6144,0.018011555075645447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,12288,0.3054257763756646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,6144,0.03568444318241543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,5120,0.015370666980743408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,6144,0.32783910963270396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,5120,0.03157688842879401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,5120,0.33151732550727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,5120,0.01069866700304879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,4096,0.013371555341614617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,4096,0.02787733409139845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,4096,0.33486755688985187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,7168,0.32114844852023655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,3584,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,6144,0.01204533295498954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,3584,0.02607022225856781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,3072,0.01109866632355584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,3584,0.3235929012298584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,3584,0.009000000026490953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,3072,0.022982221510675218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,2560,0.009720889230569204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,3072,0.008360000120268928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,3072,0.32243556446499294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,2560,0.020656888683636982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,2048,0.008647110727098253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,2560,0.008006222546100616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,2560,0.3076551225450304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,2048,0.018235555953449674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,4096,0.009377777576446533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,1536,0.006637333167923822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,2048,0.32544978459676105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,1536,0.016172443826993305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,1024,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,1536,0.007288888924651676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,1024,0.014816888504558139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,1024,0.006600889066855113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,768,0.0053635554181204895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,768,0.014150222142537435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,2048,0.007712889048788283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,768,0.32734399371676975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,768,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,512,0.004927999857399198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,1536,0.30804443359375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,512,0.013422222601042854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,512,0.31459911664326984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,512,0.006050666587220297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,256,0.004232888834344016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,1024,0.33446666929456925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,256,0.012792888614866467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,256,0.005930666708283954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,128,0.0038693332009845306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,256,0.3392186694675022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,128,0.012784889174832238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,64,0.0038453332251972626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,4096,128,0.0059573331640826324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,4096,128,0.3182480070326063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,64,0.012491555677519904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,4096,32,0.012758221891191272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,65536,0.1565528843137953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,65536,0.3032880094316271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,65536,0.07415555583106147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,65536,0.42760621176825625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,16384,0.04353955719206068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,4096,32,0.0039048890272776284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,16384,0.02202044427394867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,12288,0.0336844457520379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,16384,0.33446666929456925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,12288,0.06325599882337782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,12288,0.32296710544162327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,10240,0.028855111863878038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,12288,0.0177848885456721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,10240,0.05298133360015022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,10240,0.330639123916626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,8192,0.023730667101012334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,16384,0.08348444435331558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,8192,0.04359022114011976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,8192,0.30817778905232746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,8192,0.013883555928866068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,7168,0.03943022092183431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,7168,0.3356035550435384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,10240,0.015860444969601102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,6144,0.018018666240904067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,7168,0.012858666479587555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,6144,0.03466044531928168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,5120,0.015670221712854173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,7168,0.02145333256986406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,6144,0.011613333390818702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,6144,0.3234124448564318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,5120,0.030829333596759375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,4096,0.013914666242069669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,5120,0.010088000032636855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,5120,0.30065155029296875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,4096,0.02716355522473653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,3584,0.012605333493815528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,4096,0.009111111362775167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,4096,0.3351671165890164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,3072,0.011543110840850405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,3584,0.31328622500101727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,3072,0.02198933396074507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,3072,0.00832622249921163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,3072,0.3179786735110813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,2560,0.010287111004193624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,2560,0.020399111840460036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,2560,0.3306879997253418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,2560,0.007956444389290279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,2048,0.008816888762844933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,3584,0.025421332981851365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,2048,0.01790488925245073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,1536,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,3584,0.008671110702885522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,2048,0.007320889168315464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,2048,0.31818310419718426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,1024,0.0052577778697013855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,1536,0.0162444445821974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,1536,0.006991110742092133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,1024,0.014734221829308404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,768,0.004921777794758479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,1024,0.0063235556913746735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,1024,0.31940799289279515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,768,0.014167111780908374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,512,0.004563555535342959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,768,0.00628977765639623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,768,0.31623021761576336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,512,0.013423110875818463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,512,0.005936000082227919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,512,0.33458132214016384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,256,0.01275111072593265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,256,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,256,0.3235164483388265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,1536,0.30167465739780003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,128,0.012796444197495779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3584,128,0.2849351035224067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,64,0.0038035557501845886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3584,128,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,32,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,64,0.0124177775449223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,256,0.003954666770166821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3584,32,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,65536,0.1305262247721354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,65536,0.2724097834693061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,65536,0.07312977976269193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,16384,0.03609688745604621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3584,128,0.003536000019974179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,65536,0.4173111120859782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,16384,0.07352266709009807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,12288,0.029323554701275293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,16384,0.34169867303636337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,16384,0.02163822286658817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,12288,0.05603910817040337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,10240,0.024901333782407973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,12288,0.017621333400408428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,12288,0.3310026592678494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,10240,0.046591109699673124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,8192,0.019912888606389362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,10240,0.34794312053256565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,8192,0.03964622153176202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,8192,0.33411555820041233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,8192,0.013665777941544851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,7168,0.01793599956565433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,7168,0.03572977913750543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,7168,0.34450043572319883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,7168,0.01240888900227017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,6144,0.01575644479857551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,10240,0.016007110476493835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,6144,0.03190755512979295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,5120,0.014108444253603617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,6144,0.01125511113140318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,6144,0.33399200439453125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,5120,0.028297778632905748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,5120,0.010603555374675326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,4096,0.01219288922018475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,5120,0.3216266632080078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,4096,0.02546577817863888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,3584,0.011019555230935415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,4096,0.009031111167536842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,4096,0.34760888417561847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,3584,0.023356444305843774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,3072,0.009896888501114314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,3584,0.0087013335691558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,3584,0.3377484480539958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,3072,0.02029244436158074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,2560,0.008638222184446123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,3072,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,3072,0.3399084409077962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,2560,0.018212444252438016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,2048,0.007420444654093848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,2560,0.007707555260923173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,2560,0.33333510822719997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,2048,0.016564443707466125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,1536,0.0063573333124319715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,2048,0.007615110940403408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,2048,0.3269253306918674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,1536,0.015150222513410779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,1024,0.0052373334765434265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,1536,0.0069395556218094295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,1536,0.3278408845265706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,1024,0.014183999763594734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,768,0.004518222063779831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,1024,0.006576889091067844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,1024,0.32286045286390513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,512,0.004223111189073986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,768,0.01352266636159685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,768,0.3303297890557183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,512,0.013391999734772576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,256,0.003910222401221593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,512,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,256,0.01276622215906779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,512,0.4501484235127767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,128,0.003617777592606015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,256,0.0058755555914507965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,256,0.33416001001993817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,128,0.012430222498046027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,64,0.0032275555034478507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,3072,128,0.30660888883802623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,3072,32,0.0034106667670938703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,768,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,32,0.01236977759334776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,65536,0.1070435576968723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,65536,0.2687102158864339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,65536,0.4077671104007297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,65536,0.07155199845631917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,16384,0.03724088933732774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,3072,128,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,16384,0.07180000013775296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,16384,0.0214684440029992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,3072,64,0.012129777835475074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,12288,0.02751200066672431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,16384,0.310007996029324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,10240,0.023211555348502264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,12288,0.017510222064124215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,12288,0.3023262288835314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,10240,0.045703109767701894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,8192,0.019110222657521565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,10240,0.01552888916598426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,10240,0.33518489201863605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,8192,0.03871466716130575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,7168,0.017044444878896076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,8192,0.013177778157922955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,8192,0.3184479872385661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,7168,0.03542755709754096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,6144,0.015459555718633862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,7168,0.01146399974822998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,7168,0.332331551445855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,6144,0.031367109881507024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,12288,0.05386311146948072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,5120,0.013176889055305056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,6144,0.325546662012736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,5120,0.028095109595192805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,5120,0.3258177704281277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,5120,0.010105777945783403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,4096,0.010106667048401302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,4096,0.024360888534122046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,4096,0.009069333473841349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,3584,0.0094373333785269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,6144,0.011056000159846412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,3584,0.02202844454182519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,3584,0.00867022242810991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,3072,0.008410666551854875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,3584,0.3120195600721571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,3072,0.019655111763212416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,3072,0.32840000258551705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,3072,0.008274666965007782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,2560,0.007305777735180325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,4096,0.3318400118086073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,2560,0.018082666728231642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,2560,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,2048,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,2560,0.3235066731770833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,2048,0.016590222716331482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,1536,0.005886222339338726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,2048,0.007287999822033777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,2048,0.3335262139638265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,1536,0.015049777097172208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,1024,0.0053022222386466126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,1536,0.007075555622577667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,1536,0.3333066569434272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,1024,0.013783999615245394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,768,0.00489155575633049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,1024,0.006576889091067844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,1024,0.3386213249630398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,768,0.01348000019788742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,768,0.0063564446237352155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,768,0.3289155430263943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,512,0.013120888835854001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,512,0.006272000157170826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,512,0.3223715623219808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,256,0.003943110919660992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,256,0.012705778082211813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,256,0.3408862219916449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,128,0.003968888686762916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,128,0.012486221889654795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,512,0.004394666602214177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,64,0.003510222252872255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,128,0.005967999911970562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2560,32,0.003590222034189436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,64,0.012127111355463663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2560,128,0.442704889509413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,65536,0.09741689099205865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2560,32,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,65536,0.07007377677493624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,65536,0.2635253270467122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,16384,0.03304977880583869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2560,256,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,65536,0.4103226661682129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,16384,0.06979911194907294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,12288,0.02052622205681271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,16384,0.021049777666727703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,16384,0.3256044387817383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,10240,0.0184551113181644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,12288,0.052417778306537204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,12288,0.01700266698996226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,12288,0.30238045586480033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,8192,0.01499288943078783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,10240,0.044608000251981944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,10240,0.014837332897716098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,10240,0.3168293370140923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,7168,0.014039110806253223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,8192,0.037985778517193265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,8192,0.012468444804350534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,8192,0.3235039975908068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,6144,0.012503110700183444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,7168,0.035022222333484225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,7168,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,7168,0.31553154521518284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,5120,0.010971555279360877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,6144,0.030886222918828327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,6144,0.01072266697883606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,6144,0.32815199428134495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,5120,0.027888887458377417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,4096,0.009738666315873465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,5120,0.3325715594821506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,4096,0.023229334089491103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,3584,0.0086666668454806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,4096,0.31820445590549046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,3584,0.02183377742767334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,3584,0.00868266655339135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,3584,0.31689243846469456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,3072,0.007684444387753804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,3072,0.01956533392270406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,3072,0.28071734640333385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,5120,0.010013333625263637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,3072,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,2560,0.0069715554515520734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,2560,0.017863111363516915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,4096,0.008975110948085785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,2560,0.33790045314364964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,2048,0.0063484443558586975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,2560,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,2048,0.016615110966894362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,2048,0.007334222396214803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,2048,0.33336623509724933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,1536,0.014903111590279473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,1536,0.00692622239391009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,1024,0.004908444566859139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,1024,0.013847110999955071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,1536,0.005567111074924469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,1024,0.3281422191196018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,768,0.004328000048796336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,1024,0.006583111153708563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,768,0.013447110851605734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,1536,0.3300746546851264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,768,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,768,0.3151297834184435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,512,0.013079110946920184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,512,0.006008000009589725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,512,0.31453688939412433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,256,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,256,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,256,0.33016355832417804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,256,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,128,0.003591999825504091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,128,0.01239999963177575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,2048,128,0.30667466587490505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,2048,128,0.005607111172543631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,64,0.0035075553589397003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,32,0.0032595555401510666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,64,0.01205599970287747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,2048,32,0.012098666694429187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,65536,0.07240266932381524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,65536,0.22449334462483725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,65536,0.4066426753997803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,2048,512,0.003927999900446998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,16384,0.0247022228108512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,65536,0.06853155295054118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,12288,0.02012711101108127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,16384,0.020979555116759405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,16384,0.0587022238307529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,16384,0.31445243623521596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,12288,0.04487733377350701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,10240,0.016185777054892648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,10240,0.03813955518934462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,10240,0.3099048932393392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,8192,0.013878222141000958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,8192,0.03310577736960517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,8192,0.31210401323106557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,8192,0.012499555945396423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,7168,0.01260444439119763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,12288,0.016153777639071148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,7168,0.03093155556254917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,10240,0.014456000592973499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,6144,0.011117333339320289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,7168,0.011407111254003314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,7168,0.4198639922671848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,6144,0.027105778455734253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,5120,0.009777777724795872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,6144,0.010748444332016839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,6144,0.3219377729627821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,5120,0.023330666952663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,4096,0.008402666283978356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,5120,0.010055111514197456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,4096,0.02070755594306522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,12288,0.3351706663767497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,5120,0.37258044878641766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,3584,0.007731555236710443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,4096,0.008982222113344405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,4096,0.3178844451904297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,3072,0.007164444360468123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,3584,0.019582221905390423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,3584,0.008637333081828224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,3584,0.3265715440114339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,2560,0.006600889066855113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,3072,0.017601778109868366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,3072,0.008003555238246918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,3072,0.34362133344014484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,2560,0.016494222813182406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,2048,0.00546577779783143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,2560,0.007673777639865875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,2560,0.31862399313184947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,1536,0.0052382221652401825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,2048,0.015483554866578845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,2048,0.007330666813585494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,2048,0.3355093267228868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,1536,0.014150222142537435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,1536,0.006924444602595435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,1024,0.004272888931963178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,1536,0.32344177034166127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,768,0.003963555726740095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,1024,0.006530666516886816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,1024,0.3329182200961643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,768,0.013096888860066732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,512,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,768,0.006276444428496891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,768,0.34242134624057347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,512,0.01274222218328052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,512,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,512,0.3268444538116455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,256,0.0032017777363459268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,256,0.012073777616024017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,1024,0.01350044459104538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,256,0.34467466672261554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,128,0.0032239999208185407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,256,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,128,0.012094222009181976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,64,0.0028906667398081887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1536,128,0.2982737753126356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1536,32,0.003048000029391713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,64,0.011701333026091257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,65536,0.055719110700819224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,65536,0.2224444548288981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1536,128,0.0058746664888328975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,65536,0.411772436565823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,16384,0.018200000127156574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,65536,0.06686933173073663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1536,32,0.011769777370823754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,16384,0.05713866816626655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,12288,0.015828443898095023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,16384,0.01961511042382982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,12288,0.04353422257635328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,12288,0.34368620978461367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,10240,0.01423288881778717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,10240,0.03762844536039565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,10240,0.3482186794281006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,16384,0.33386754989624023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,8192,0.013434666726324292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,10240,0.013776889277829064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,8192,0.031806220610936485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,12288,0.015269334117571512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,7168,0.012183999849690331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,8192,0.012101333174440594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,7168,0.02940711047914293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,6144,0.010568888651000129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,7168,0.30943287743462455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,6144,0.02574577762020959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,6144,0.31405867470635307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,5120,0.009324444664849175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,6144,0.010779555473062726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,5120,0.023058666123284235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,8192,0.3079555564456516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,4096,0.007642666498819987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,5120,0.39136799176534015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,7168,0.011356444822417365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,4096,0.020671111014154222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,3584,0.0075048887067370946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,4096,0.008983111215962304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,4096,0.29904177453782826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,3584,0.019239111079110038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,3072,0.007101332975758447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,3584,0.008676444490750631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,3584,0.33065954844156903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,3072,0.01752000053723653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,2560,0.006245333287451003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,3072,0.008031999899281396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,3072,0.3162088923984104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,5120,0.010020444790522257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,2560,0.01625155574745602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,2048,0.005679111099905438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,2560,0.00790666706032223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,2560,0.32748622364468044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,2048,0.015213333898120456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,1536,0.005241777747869492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,2048,0.007293333609898885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,2048,0.303455114364624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,1536,0.014140443669425117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,1024,0.00452533322903845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,1536,0.006979555719428592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,1536,0.3070613278283013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,768,0.004232888834344016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,1024,0.00655377780397733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,1024,0.31020355224609375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,768,0.012706666356987424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,512,0.00387999994887246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,768,0.006254222244024277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,768,0.31908710797627765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,512,0.012400888734393649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,256,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,512,0.00598933340774642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,512,0.31951644685533315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,256,0.0120328888297081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,128,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,256,0.005967999911970562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,256,0.324139568540785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,128,0.012073777616024017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,64,0.003212444484233856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,1024,128,0.0058906666106647914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,1024,0.01309511148267322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,1024,32,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,1024,128,0.29704533682929146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,65536,0.04154133465554979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,64,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,1024,32,0.011695110963450538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,16384,0.015000889698664347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,65536,0.0649626652399699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,65536,0.2216115527682834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,16384,0.05583110782835218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,65536,0.4068133301205105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,12288,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,16384,0.018952000472280715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,16384,0.33090045717027455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,12288,0.042585776911841504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,10240,0.01312800000111262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,12288,0.01516711049609714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,12288,0.33847732014126247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,10240,0.03672444489267137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,8192,0.012792888614866467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,10240,0.013824000126785703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,10240,0.33444976806640625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,8192,0.03088444471359253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,8192,0.01187999960449007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,7168,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,8192,0.3368382189008925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,7168,0.0288702216413286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,6144,0.01032977799574534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,7168,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,7168,0.3433377742767334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,6144,0.025674665967623394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,6144,0.010377777947319878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,6144,0.3454586664835612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,5120,0.022426666484938726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,5120,0.3230951097276476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,5120,0.010033778018421596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,4096,0.007992000215583378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,4096,0.0199306673473782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,4096,0.32366755273607045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,4096,0.008988444175985124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,3584,0.007270221908887227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,5120,0.008970666262838576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,3584,0.018933332628673978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,3072,0.006393777827421824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,3584,0.3181706534491645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,3584,0.00869777798652649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,3072,0.017215111189418368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,2560,0.005943999936183293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,3072,0.32141245736016166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,2560,0.016169778174824186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,2560,0.007671111159854465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,2560,0.3224968910217285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,2048,0.0053048887186580235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,2048,0.015159999330838522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,2048,0.3421786626180013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,2048,0.007325333025720384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,1536,0.004916444420814514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,3072,0.008019555773999956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,1536,0.013849777479966482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,1024,0.0041662222809261745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,1536,0.3265235424041748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,1536,0.006976889239417181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,768,0.0041831110914548235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,1024,0.013076444466908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,1024,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,768,0.012729778057999082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,512,0.003552000141806073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,768,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,1024,0.42568267716301816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,768,0.3281351195441352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,256,0.0032293332947625052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,512,0.005958222266700532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,512,0.33145509825812447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,256,0.012048888537618848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,128,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,256,0.0058986664646201665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,128,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,256,0.38336533970303005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,64,0.0032142222755485107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,768,128,0.0058959999846087555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,768,128,0.30014310942755806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,768,32,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,65536,0.03197244471973843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,64,0.01204711116022534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,32,0.011736888852384357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,768,512,0.012473777764373355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,65536,0.22052444352043998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,65536,0.06308444341023763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,65536,0.4029982089996338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,16384,0.054711110062069364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,16384,0.018607111440764535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,16384,0.3319182131025526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,12288,0.043603556023703686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,12288,0.32805689175923664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,10240,0.010407999985747868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,10240,0.03637689020898607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,16384,0.013197333448463015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,10240,0.013758222262064615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,8192,0.00906044410334693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,10240,0.32848887973361546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,8192,0.030894223186704848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,8192,0.32707733578152126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,7168,0.008981333010726506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,8192,0.01182577759027481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,12288,0.015150222513410779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,7168,0.011402666568756104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,7168,0.02817244331041972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,6144,0.0083137775460879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,7168,0.33236977789137095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,6144,0.02541422181659275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,6144,0.34171377287970645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,6144,0.010728889041476779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,12288,0.01182755579551061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,5120,0.00869688888390859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,5120,0.022384888596004907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,4096,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,5120,0.3265484439002143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,4096,0.019928889142142404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,4096,0.009108444054921469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,3584,0.0069431112044387394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,3584,0.01902933253182305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,5120,0.009957333405812582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,3584,0.3471333450741238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,3072,0.006311111152172089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,3584,0.008638222184446123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,3072,0.01721600029203627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,4096,0.33623645040724015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,2560,0.005877333382765452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,3072,0.34270665380689835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,3072,0.008023111356629265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,2560,0.016223111086421542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,2560,0.007627555893527136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,2048,0.005199111170238919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,2560,0.3225039905971951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,2048,0.014814222852389017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,1536,0.004580444345871608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,2048,0.007284444239404466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,2048,0.3319635656144884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,1536,0.013473778135246701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,1536,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,1536,0.3309475580851237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,1024,0.012998222476906247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,1024,0.33339911037021214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,768,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,768,0.012735111018021902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,1024,0.004254222330119875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,768,0.3347315523359511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,512,0.0037164443896876443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,768,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,512,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,1024,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,512,0.005986666513813867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,256,0.003552888830502828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,512,0.32160178820292157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,256,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,256,0.005636444522274866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,128,0.0032382222513357797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,256,0.3395822313096788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,128,0.011881777809725868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,64,0.0031831111345026228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,512,32,0.0029306666304667792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,512,128,0.005715555614895291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,512,128,0.29748355017768013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,65536,0.022169777088695105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,64,0.011744000017642975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,512,32,0.011700444751315646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,16384,0.01204711116022534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,65536,0.061420447296566434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,65536,0.21967021624247232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,16384,0.05407555566893684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,12288,0.010151111417346531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,65536,0.43062665727403426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,16384,0.34070756700303817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,12288,0.041638222005632185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,10240,0.009711110757456886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,12288,0.3103768825531006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,12288,0.015122666954994202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,10240,0.03630222214592828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,10240,0.013788444300492605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,8192,0.00868800034125646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,10240,0.30811911159091526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,8192,0.030995554394192163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,7168,0.008384000096056197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,8192,0.0120319997270902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,8192,0.32986754841274685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,7168,0.02821422285503811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,16384,0.0181697772608863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,6144,0.007968000239796108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,7168,0.011247110863526663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,7168,0.3290346728430854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,5120,0.008100444244013892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,6144,0.02475999957985348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,6144,0.33611467149522567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,5120,0.02260355485810174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,4096,0.007671111159854465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,5120,0.31924711333380806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,4096,0.020120888948440552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,4096,0.008935111264387766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,4096,0.31759466065300834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,3584,0.006919999917348226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,3584,0.018778666853904724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,3584,0.33355021476745605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,6144,0.010675555302037133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,3072,0.006315555423498154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,3072,0.01717866626050737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,5120,0.009888889061080085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,3072,0.323928885989719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,3072,0.008007999923494127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,2560,0.016176000237464905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,2560,0.34177777502271867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,2560,0.007681777907742395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,2048,0.0052844443255000645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,3584,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,2048,0.014898666077189975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,2048,0.3202657699584961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,2048,0.007366221812036302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,1536,0.004615111069546806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,1536,0.013751111096805997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,1536,0.3260844548543294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,1024,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,1024,0.012759110993809171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,1024,0.3408239947424994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,2560,0.0058622223635514574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,768,0.0038924444880750445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,1024,0.006342222293217977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,768,0.01276088919904497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,1536,0.0069848886794514125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,768,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,512,0.003499555504984326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,768,0.32290755377875435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,512,0.012057777908113269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,256,0.003157333367400699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,512,0.3457289006974962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,256,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,256,0.005575110928879843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,128,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,128,0.011971555650234222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,512,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,128,0.3196764522128635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,64,0.0028693332440323303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,256,128,0.0058959999846087555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,256,32,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,64,0.011391999820868174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,65536,0.019021333919631112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,256,32,0.01145600030819575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,256,256,0.33755821651882595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,16384,0.009746666583749983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,65536,0.21951288647121855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,65536,0.0580728915002611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,65536,0.4060595565372043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,12288,0.008627555436558193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,16384,0.0181377778450648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,16384,0.3408186700608995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,12288,0.040680001179377236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,10240,0.008994667066468133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,12288,0.015133332875039844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,12288,0.3402426772647434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,10240,0.03501688771777683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,8192,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,10240,0.32052800390455455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,8192,0.029880887932247583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,8192,0.32905599806043834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,8192,0.012155555188655853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,7168,0.007970666719807519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,16384,0.05320088730918037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,7168,0.02755555510520935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,6144,0.007672000262472365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,7168,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,7168,0.3250044451819526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,6144,0.024715556038750544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,5120,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,6144,0.010632889138327705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,10240,0.013521778086821238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,5120,0.02237333357334137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,6144,0.4091155529022217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,4096,0.007291555404663086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,5120,0.00998133338159985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,5120,0.317047119140625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,3584,0.007272889216740926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,4096,0.008952000074916417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,4096,0.32745244767930776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,3584,0.019156444403860304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,3072,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,3584,0.008392000363932716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,3584,0.329219553205702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,3072,0.01732711162832048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,2560,0.005581333405441708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,3072,0.007995555798212687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,3072,0.31777776612175834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,2560,0.016210666961140104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,2048,0.005158222383923001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,2560,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,2560,0.33342933654785156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,2048,0.015171556009186639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,1536,0.004552888787455029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,4096,0.020184889435768127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,2048,0.007369777394665613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,2048,0.3195519977145725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,1024,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,1536,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,1536,0.3064195579952664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,1024,0.012464889221721225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,768,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,1024,0.006280888699822956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,1024,0.3285280068715413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,768,0.01220355596807268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,512,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,768,0.0063182223174307085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,768,0.3224257893032498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,512,0.012061333490742577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,512,0.33328978220621747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,512,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,256,0.0031964443624019623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,1536,0.013822221921549903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,256,0.011715555356608497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,128,0.0028684443483750024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,256,0.005778666585683823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,256,0.3286142349243164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,128,0.01165600038237042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,64,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,384,128,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,128,32,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,384,128,128,0.32319021224975586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,64,0.011415999796655444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,128,32,0.01149777736928728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,16384,0.009424000150627559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,65536,0.21865688429938424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,16384,0.05250222153133816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,12288,0.04046666622161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,10240,0.009023110899660323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,10240,0.035162667433420815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,8192,0.00795022232664956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,65536,0.014492443866199918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,7168,0.007618666523032718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,8192,0.03018933203485277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,6144,0.007287999822033777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,7168,0.028429332706663344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,5120,0.0069422221018208405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,12288,0.008330666356616551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,4096,0.006614222294754452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,6144,0.025203555822372437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,3584,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,5120,0.0225564440091451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,3072,0.0063440000845326324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,4096,0.019871999820073444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,2560,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,3584,0.018992889258596633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,2048,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,3072,0.017198221551047433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,1536,0.004559999952713649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,2560,0.01550222271018558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,1024,0.0038986665507157645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,2048,0.01482488877243466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,768,0.0035706667436493766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,1536,0.013776000175211163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,512,0.003236444460021125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,1024,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,256,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,768,0.012447111308574677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,128,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,512,0.012081777883900536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,64,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,256,0.011387555963463254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,64,32,0.0028328889360030494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,128,0.011720889144473605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,65536,0.0169813334941864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,64,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,16384,0.008970666262838576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,64,32,0.011063999599880643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,12288,0.008391111261314815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,16384,0.05217866765128242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,65536,0.2185751067267524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,10240,0.007983110845088959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,12288,0.03976533479160733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,8192,0.00831822223133511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,7168,0.007643555601437886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,10240,0.03526399864090814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,8192,0.029888000753190782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,6144,0.007316444483068254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,5120,0.00702400008837382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,4096,0.006734222173690796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,6144,0.02473066747188568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,3584,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,4096,0.020304888486862183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,3584,0.018564444449212816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,3072,0.006311999840868845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,2560,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,3072,0.01684977776474423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,2048,0.006229333165619109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,2560,0.016215999921162922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,1536,0.004516444272465176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,2048,0.014450665977266101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,1024,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,1536,0.01350400017367469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,768,0.003544888976547453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,7168,0.028063111835055884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,512,0.0031795555518733132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,768,0.0124195557501581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,256,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,512,0.01221155540810691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,128,0.0029013332807355454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,256,0.011752888560295105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,64,0.0028088889602157804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,128,0.011911999848153857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,384,32,32,0.002546666603949335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,64,0.01146399974822998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,32,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,16384,0.49172268973456484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,1024,0.012416000167528788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,384,32,5120,0.02253866692384084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,16384,0.36621689796447754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,12288,0.37351377805074054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,16384,0.13471466965145534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,12288,0.27592089441087514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,12288,0.09848444329367743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,16384,0.3742746777004666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,10240,0.2498533460828993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,12288,0.35030399428473574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,10240,0.23490667343139648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,8192,0.20378400219811332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,10240,0.08388355705473159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,8192,0.19050578276316324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,10240,0.32157333691914874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,7168,0.18059644434187147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,8192,0.07247644662857056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,8192,0.3426542282104492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,7168,0.17011555035909018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,6144,0.15683644347720677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,7168,0.06102755334642199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,6144,0.1455155611038208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,7168,0.3285271061791314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,5120,0.13199822107950845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,6144,0.054676443338394165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,5120,0.12485244539048936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,6144,0.3357626597086589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,4096,0.10593599743313259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,5120,0.04643377661705017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,4096,0.10041244162453546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,5120,0.33737866083780926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,4096,0.038910223378075495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,3584,0.09042488866382176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,3584,0.09155555566151936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,4096,0.3298248979780409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,3072,0.079748445087009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,3584,0.33064977327982586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,3072,0.07939644654591878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,2560,0.06783911254670885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,3072,0.03146133489078946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,3072,0.3465217749277751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,2560,0.06846488846672906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,2048,0.05949422385957506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,2560,0.028279999891916912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,2560,0.29393066300286186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,2048,0.05863733424080742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,1536,0.04351644383536445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,3584,0.03472266594568888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,2048,0.02492800023820665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,2048,0.3370631006028917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,1536,0.04799555407630073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,1024,0.03187733226352268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,1536,0.01925511161486308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,1536,0.29833687676323783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,1024,0.03935199975967407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,1024,0.01481244464715322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,768,0.023208000593715247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,1024,0.32031554645962185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,768,0.033589333295822144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,768,0.013235555754767524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,768,0.3303786648644341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,512,0.028218666712443035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,256,0.013078221844302284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,512,0.012479110724396177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,512,0.31620976659986705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,256,0.02535999980237749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,256,0.011167999770906238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,256,0.33347288767496747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,128,0.02446488870514764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,65536,128,0.011077333655622272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,65536,128,0.3038506772783068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,64,0.01072088877360026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,32,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,64,0.023706666297382776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,65536,32,0.0298977792263031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,65536,0.3766213258107503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,128,0.01073333372672399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,65536,0.4457893371582031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,16384,0.10062933630413479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,65536,512,0.017690666847758822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,65536,0.14465510845184326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,16384,0.12083644337124294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,12288,0.07838577694363065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,16384,0.042713777886496655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,12288,0.09329155418607925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,16384,0.33543554941813153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,10240,0.06748177607854207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,12288,0.33432711495293516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,10240,0.08058844672309028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,65536,0.5082097583346897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,10240,0.3352968957689073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,8192,0.05542311072349548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,8192,0.06647821929719713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,8192,0.025427555044492085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,8192,0.3263546625773112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,7168,0.048319111267725624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,12288,0.03312266535229153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,7168,0.021465778350830078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,6144,0.04281511240535312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,10240,0.029353777567545574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,7168,0.33392977714538574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,6144,0.05155200097295973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,6144,0.01917777789963616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,6144,0.33691022131178117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,5120,0.03629066546758016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,5120,0.04743466774622599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,5120,0.3439173433515761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,5120,0.017222222354676988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,7168,0.059154669443766274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,4096,0.029935999049080744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,4096,0.03875644339455499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,4096,0.014840000205569796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,3584,0.027458666099442378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,4096,0.31861866845024955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,3584,0.034494221210479736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,3072,0.023585778143670823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,3584,0.3376302189297146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,3072,0.03107200066248576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,3072,0.3363022274441189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,2560,0.020928889513015747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,2560,0.027540443672074214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,3584,0.013596444494194455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,2560,0.3473715517255995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,2048,0.017117333081033494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,2560,0.011745777395036487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,2048,0.024586665961477492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,3072,0.01258399999803967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,1536,0.013747555514176687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,2048,0.01072266697883606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,2048,0.3331555525461833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,1536,0.021262221866183813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,1536,0.008681777450773451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,1024,0.010798222488827176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,1536,0.33036088943481445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,1024,0.019038221902317472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,768,0.009501333038012186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,1024,0.007672000262472365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,1024,0.3310515615675184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,768,0.017117333081033494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,512,0.007824000385072496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,768,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,768,0.3307182259029812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,512,0.01588711142539978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,256,0.0052373334765434265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,512,0.006963555183675554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,512,0.33444799317253965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,128,0.00498844434817632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,256,0.015171556009186639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,256,0.006693333387374878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,128,0.014505777094099255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,256,0.31819110446506077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,64,0.00489688871635331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,16384,32,0.006344888773229387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,16384,128,0.30676621860928005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,16384,128,0.006799110935793982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,64,0.01477777792347802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,16384,32,0.016034666034910414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,65536,0.25324177742004395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,16384,0.07156088617112902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,65536,0.386791123284234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,65536,0.13572177622053358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,16384,0.10791199737124973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,12288,0.05679555733998617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,16384,0.04001511136690775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,12288,0.08365688721338908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,12288,0.030809776650534734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,12288,0.3128506607479519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,10240,0.048787554105122886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,65536,0.4928604231940375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,10240,0.07072177860471937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,8192,0.0396382212638855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,10240,0.02686311139000787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,10240,0.34354933102925617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,8192,0.05873955620659722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,7168,0.03496177660094367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,8192,0.022952889402707417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,16384,0.3048906591203478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,8192,0.32514932420518666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,6144,0.0306951105594635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,7168,0.0518853333261278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,7168,0.019860444797409903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,7168,0.35702045758565265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,6144,0.04815466536415947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,5120,0.026678222748968337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,6144,0.01791288952032725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,6144,0.3185093402862549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,5120,0.04162044326464335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,5120,0.016203555795881484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,5120,0.32623288366529674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,4096,0.0342497792508867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,3584,0.020040000478426617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,4096,0.014133334159851074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,4096,0.33922754393683535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,3584,0.030202666918436687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,3072,0.01744622157679664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,3584,0.01313066648112403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,3584,0.3162124421861437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,3072,0.02772533396879832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,2560,0.015318221516079374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,3072,0.012324444121784635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,3072,0.32704710960388184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,2560,0.02530844509601593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,4096,0.022894221875402663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,2048,0.014479110638300577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,2560,0.328761789533827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,2560,0.011065777805116443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,2048,0.022209778428077698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,1536,0.01074399964676963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,2048,0.00960799968904919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,2048,0.32139645682440865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,1536,0.0196426659822464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,1024,0.008804444637563493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,1536,0.008344000412358178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,1024,0.017268444101015728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,1024,0.007284444239404466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,1024,0.3552195496029324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,768,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,768,0.015916445189052157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,768,0.3471973207261827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,512,0.006551110910044775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,512,0.014808000789748298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,1536,0.3420497841305203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,512,0.0069475554757648045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,256,0.005225777626037598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,512,0.3174000051286486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,256,0.014527110589875115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,256,0.3226284450954861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,256,0.006646222124497096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,128,0.004858666823969947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,768,0.007279111279381647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,128,0.014179555906189812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,64,0.004556444370084339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,12288,128,0.0068604445291890045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,12288,128,0.30591821670532227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,12288,32,0.004948444250557157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,64,0.013850666582584381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,12288,32,0.014910222755538093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,65536,0.2203084495332506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,65536,0.34385334120856387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,16384,0.06158133347829183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,65536,0.1349066628350152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,16384,0.09728888670603435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,12288,0.048179556926091514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,16384,0.31596801016065806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,16384,0.039028445879618325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,12288,0.0746142201953464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,10240,0.04202399982346428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,65536,0.48808977339002824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,12288,0.029887109994888306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,10240,0.06486666864818998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,8192,0.03407377666897244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,10240,0.025693333811230127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,10240,0.3181048764122857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,8192,0.05308711197641161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,7168,0.030279109875361126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,8192,0.022023111581802368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,8192,0.40714666578504777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,7168,0.0471111105548011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,6144,0.026825777358478967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,7168,0.019394665956497192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,7168,0.3218337694803874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,6144,0.04234310984611511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,5120,0.023758222659428913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,6144,0.017498667041460674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,12288,0.3205617798699273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,6144,0.2939564387003581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,5120,0.03702755437956916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,5120,0.01594577729701996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,5120,0.3271262115902371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,4096,0.030905776553683813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,3584,0.017229333519935608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,4096,0.013830222189426422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,4096,0.3246817853715685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,3584,0.02784088916248745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,3072,0.015262222952312894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,3584,0.012576888832781049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,3584,0.3128506607479519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,3072,0.025570667452282373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,2560,0.013530666629473368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,3072,0.011544889046086205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,3072,0.32955911424424916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,2560,0.023648000425762598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,2048,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,2560,0.010666666759385003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,2560,0.31439020898607045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,2048,0.021103999680942957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,1536,0.010111110905806223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,2048,0.31060179074605304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,2048,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,1536,0.01888799998495314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,1024,0.008319999608728621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,1536,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,1536,0.30836354361640084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,1024,0.016531555189026725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,768,0.006883555402358373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,1024,0.007307555940416124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,1024,0.28374844127231175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,768,0.015491555134455362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,768,0.007024889190991719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,768,0.3220951027340359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,512,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,512,0.3281911214192708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,512,0.0069146665434042616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,256,0.0046160001721647055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,256,0.01387466655837165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,4096,0.019762666689025033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,512,0.005250666704442766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,128,0.004457777573002709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,256,0.006682666639486949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,256,0.34225776460435653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,128,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,64,0.0038817777401871155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,10240,128,0.006583999842405319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,10240,128,0.30797600746154785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,10240,32,0.004543999830881755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,64,0.01350844403107961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,65536,0.19627200232611763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,10240,32,0.014781332678265043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,65536,0.29609155654907227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,16384,0.05376266770892673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,65536,0.08176177740097046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,16384,0.08388177553812663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,16384,0.33828266461690265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,16384,0.025721778472264607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,12288,0.04383377896414863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,65536,0.3568062252468533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,12288,0.06636889113320245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,12288,0.01964000033007728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,10240,0.036933332681655884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,12288,0.30134399731953937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,10240,0.05669689178466797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,8192,0.031146665414174397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,10240,0.017412443955739338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,10240,0.34294488694932723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,8192,0.04691911074850294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,7168,0.027299554811583623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,8192,0.015824000040690105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,8192,0.324343999226888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,7168,0.042027556233935885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,7168,0.014223110344674854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,6144,0.02479733361138238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,7168,0.3276746537950304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,6144,0.03738933470514085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,5120,0.020385776956876118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,6144,0.0129120002190272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,6144,0.329801771375868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,5120,0.032594667540656194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,5120,0.011952888634469775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,4096,0.017281777328915067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,5120,0.32510044839647084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,4096,0.02741866641574436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,4096,0.01035288886891471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,4096,0.3317715591854519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,3584,0.025404444999165002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,3072,0.01368533323208491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,3072,0.023745778534147475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,3584,0.4750969145033095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,3072,0.008745777938101027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,2560,0.012081777883900536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,3072,0.33422666125827366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,3584,0.015375999940766228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,2560,0.021564443906148274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,2048,0.01129955550034841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,2560,0.323999113506741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,2560,0.007831999825106727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,3584,0.009318222602208456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,2048,0.019324445062213473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,2048,0.007349333001507654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,2048,0.3183048831091987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,1536,0.016881777180565726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,1024,0.006613333192136552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,1536,0.329839997821384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,1024,0.015236443943447538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,1024,0.32818934652540416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,1024,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,768,0.014543111125628153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,1536,0.008606221940782335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,768,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,512,0.00489155575633049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,768,0.3400533464219835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,1536,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,512,0.014099554883109199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,512,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,256,0.0042462220622433555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,512,0.3226257695092095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,256,0.013427555561065674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,128,0.003917333152559069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,768,0.005570666657553778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,256,0.00574933323595259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,256,0.32026399506462944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,64,0.003567999849716822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,128,0.013100444442696042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,8192,128,0.006225777582989798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,8192,32,0.004250666747490565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,8192,128,0.29711911413404674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,64,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,65536,0.17310933272043863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,8192,32,0.013842666314707862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,16384,0.047464887301127114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,65536,0.2641751236385769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,65536,0.07967733012305366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,16384,0.0740728908114963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,16384,0.02408711115519206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,16384,0.3234577708774143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,12288,0.03823555509249369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,65536,0.3431306680043538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,12288,0.05840266413158841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,10240,0.032646222247017756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,12288,0.3332444561852349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,10240,0.05039111110899183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,10240,0.33391555150349933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,8192,0.02690222197108799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,8192,0.04102755586306254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,12288,0.018731556004948087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,8192,0.32049332724677193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,7168,0.023918223049905565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,8192,0.015223999818166098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,10240,0.016963554753197562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,7168,0.32750222418043345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,6144,0.021789333886570398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,7168,0.013597333596812354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,6144,0.03309244579739041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,6144,0.3399013413323297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,6144,0.012755555411179861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,5120,0.018416888184017606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,5120,0.02941333254178365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,5120,0.33080710305107963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,7168,0.03684355484114753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,4096,0.0159262220064799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,4096,0.024665777881940205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,4096,0.310331556532118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,3584,0.014104889498816596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,4096,0.009723555710580613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,3584,0.02331200076474084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,3584,0.33184888627794057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,3072,0.012297777665985955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,5120,0.011757333245542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,3072,0.02137600051032172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,3072,0.008658666577604082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,2560,0.011056889262464313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,3072,0.3100959989759657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,2560,0.019662222928471036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,2048,0.010495999621020423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,2560,0.30527644687228733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,3584,0.00903555585278405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,2048,0.017603556315104168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,1536,0.007744889292452071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,2048,0.007291555404663086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,1536,0.016194666425387066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,1536,0.0069146665434042616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,1536,0.324327998691135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,1024,0.0058462222417195635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,1024,0.014901333385043673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,1024,0.3374648888905843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,2560,0.007744889292452071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,1024,0.006575999988449945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,768,0.004913777940803104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,768,0.006264888991912206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,768,0.01424800025092231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,2048,0.302495108710395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,768,0.32682400279574925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,512,0.013834666874673633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,512,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,256,0.003949333396222857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,512,0.3043999936845568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,256,0.01313688854376475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,128,0.003647110942337248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,256,0.005949333310127258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,256,0.39664443333943683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,7168,128,0.30199021763271755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,64,0.003504000190231535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,512,0.004221333397759331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,7168,32,0.0038559999730851916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,64,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,32,0.013408000270525614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,65536,0.16007288297017416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,65536,0.2529928949144152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,7168,128,0.012826667063766055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,16384,0.045829332537121244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,7168,128,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,65536,0.07670133643680148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,65536,0.33774222267998594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,16384,0.07134399811426799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,12288,0.03570222192340427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,16384,0.022895110978020564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,12288,0.05704444646835327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,12288,0.01814133259985182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,16384,0.3673502339257135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,12288,0.30664266480339897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,10240,0.04818222257826063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,8192,0.025399110383457605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,10240,0.01643466618325975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,10240,0.3499991099039714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,8192,0.03961688942379422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,8192,0.014682667122946845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,7168,0.02217866645918952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,8192,0.3275795512729221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,7168,0.0363031095928616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,6144,0.01976355579164293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,7168,0.01312444441848331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,7168,0.31698044141133624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,10240,0.030866665972603693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,6144,0.03225955698225234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,5120,0.016959110895792644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,6144,0.33416001001993817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,6144,0.012254222399658628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,5120,0.029000888268152874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,4096,0.014920888675583733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,5120,0.011065777805116443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,5120,0.328384002049764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,4096,0.024724443753560383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,3584,0.012960000170601739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,4096,0.009365333451165093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,4096,0.33590488963656956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,3584,0.023064000738991633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,3584,0.00870222184393141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,3072,0.011716444459226398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,3584,0.3370506763458252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,3072,0.021320889393488567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,2560,0.010309333602587381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,3072,0.008392000363932716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,3072,0.3368346691131592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,2560,0.01969688965214623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,2560,0.007627555893527136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,2560,0.34059466256035703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,2048,0.01793066660563151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,2048,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,2048,0.33256178432040745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,1536,0.008097777764002482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,1536,0.015790222419632804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,1536,0.29676800303988987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,1024,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,1536,0.00693866651919153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,2048,0.00959911114639706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,1024,0.014504889647165934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,768,0.005579555614127054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,1024,0.006331555545330048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,1024,0.33728623390197754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,768,0.014113777213626437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,512,0.004932444542646408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,768,0.0063457778758472866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,768,0.33334488338894314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,512,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,256,0.004285333471165763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,512,0.3388693332672119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,256,0.013049778011110095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,256,0.005944889038801193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,256,0.3255191114213732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,128,0.0038933331767717996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,128,0.0127368892232577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,6144,128,0.31131111250983345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,128,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,64,0.00388355553150177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,6144,32,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,64,0.012485333614879184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,6144,512,0.005943999936183293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,6144,32,0.013526221944226159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,65536,0.13022043969896105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,16384,0.03676266802681817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,65536,0.2526720099978977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,65536,0.07380088832643297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,16384,0.07123555739720662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,12288,0.028985778490702312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,16384,0.022088888618681166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,65536,0.3506693310207791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,16384,0.3351893424987793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,10240,0.025961778230137292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,12288,0.054880887269973755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,12288,0.01769244505299462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,12288,0.33867822753058535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,10240,0.04653244548373752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,10240,0.0158933334880405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,8192,0.03891911109288534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,10240,0.3997137811448839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,7168,0.018516444497638278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,8192,0.014205333259370593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,8192,0.3311048878563775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,7168,0.03470933437347412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,7168,0.013074666261672974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,6144,0.01683911184469859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,7168,0.31444178687201607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,6144,0.03171733352873061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,5120,0.014239999983045789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,6144,0.012033777932325998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,6144,0.31141334109836155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,8192,0.022188444932301838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,5120,0.028622223271263972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,5120,0.010422222316265106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,5120,0.31795377201504177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,4096,0.02437688906987508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,3584,0.011209777659840055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,4096,0.009196444518036312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,4096,0.3173448774549696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,3584,0.02273422148492601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,3072,0.01015733347998725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,3584,0.008669333325492011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,3584,0.3249937693277995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,3072,0.02128977742460039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,2560,0.00923466682434082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,3072,0.008000888758235509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,3072,0.3188604513804118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,2560,0.018954666124449838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,4096,0.01314933349688848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,2048,0.008803555534945594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,2560,0.33135912153455943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,2560,0.007610666255156199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,2048,0.01687022215790219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,1536,0.007088888850477006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,2048,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,2048,0.31344887945387095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,1536,0.015541333291265698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,1024,0.005548444473081165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,1536,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,1536,0.34608888626098633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,1024,0.014476444986131458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,768,0.005023111071851519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,1024,0.006454222318198945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,1024,0.32953956392076283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,768,0.013725333743625216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,512,0.004599110947714912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,768,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,768,0.33657333585951066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,512,0.013440888788965015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,256,0.004242666479614046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,512,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,512,0.33157867855495876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,256,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,128,0.0038284444146686127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,256,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,256,0.32762400309244794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,128,0.01272266689274046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,64,0.003565333369705412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,5120,128,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,5120,32,0.003540444291300244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,5120,128,0.2986240122053358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,64,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,65536,0.10694311062494914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,5120,32,0.01311644415060679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,65536,0.21186489529079863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,65536,0.07112089130613539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,16384,0.032518221272362605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,16384,0.060991108417510986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,65536,0.3338026735517714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,12288,0.026151110728581745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,16384,0.3074337906307644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,16384,0.021225778592957392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,12288,0.046762665112813316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,12288,0.01716444392999013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,10240,0.02297244469324748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,12288,0.31453333960639107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,10240,0.039573334985309176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,10240,0.015636444091796875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,10240,0.3320017655690511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,8192,0.03388355506791009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,8192,0.013482666677898832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,7168,0.016573333077960543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,7168,0.030596445004145306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,7168,0.012087110843923358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,6144,0.015228443675571017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,8192,0.019135110908084445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,7168,0.4823315408494737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,6144,0.3443422317504883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,8192,0.323832008573744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,5120,0.012924444344308643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,5120,0.02572444412443373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,5120,0.3335439893934462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,4096,0.011881777809725868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,4096,0.02176977859603034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,6144,0.02808177802297804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,4096,0.00865955568022198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,4096,0.3397875626881917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,6144,0.011023110813564725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,3584,0.010148444109492833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,3584,0.02036088870631324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,3584,0.008339555727110969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,3072,0.009837333526876237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,3584,0.2975875536600749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,3072,0.018566222654448617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,5120,0.010048000348938836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,3072,0.007978666987684038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,3072,0.32749509811401367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,2560,0.016919111212094624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,2048,0.00798044436507755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,2560,0.007638221813572778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,2560,0.3127502335442437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,2048,0.01552977826860216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,1536,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,2048,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,2048,0.36548354890611434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,1536,0.014474666780895658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,1024,0.005199111170238919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,1536,0.0069360000391801195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,1536,0.32240621248881024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,1024,0.013800000150998434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,768,0.004583111239804162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,2560,0.008427555362383524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,1024,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,1024,0.32751554912990993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,768,0.013112000293201871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,512,0.004252444538805219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,768,0.006263999889294307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,768,0.30683912171257866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,512,0.013069333301650153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,256,0.003910222401221593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,512,0.005974222388532426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,512,0.3290151225195991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,256,0.0124177775449223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,128,0.0035751110149754416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,256,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,256,0.3226204448276096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,128,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,64,0.0035182221068276297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,4096,128,0.3123973210652669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,4096,32,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,64,0.012357333468066322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,65536,0.09981777932908799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,4096,32,0.01275288893116845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,65536,0.20750310685899523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,65536,0.35129422611660427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,16384,0.029458665185504492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,4096,128,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,16384,0.05932977464463976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,16384,0.3354960017734104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,16384,0.020831111404630873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,12288,0.02316355539692773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,12288,0.04562311040030586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,12288,0.33541422420077854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,12288,0.016872889465755887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,10240,0.03968177901373969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,10240,0.346635553571913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,10240,0.015392000476519266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,8192,0.016910221841600206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,65536,0.07018666797214083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,8192,0.033193777004877724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,8192,0.013086222112178802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,7168,0.01479644411140018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,8192,0.33702577484978574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,7168,0.030414223670959473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,6144,0.013571555415789286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,7168,0.3213333288828532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,7168,0.011904888682895236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,6144,0.02776088813940684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,5120,0.011779555843936073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,6144,0.010410666465759277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,6144,0.3142666551801893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,10240,0.02049511174360911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,5120,0.0251555558707979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,4096,0.01055022246307797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,5120,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,4096,0.02128177715672387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,3584,0.009375111096435124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,4096,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,5120,0.4192817740970188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,4096,0.3219173219468859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,3072,0.008375110725561777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,3584,0.019913777709007263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,3584,0.008384000096056197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,3072,0.018758222460746765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,3584,0.31781334347195095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,2560,0.007631111476156447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,3072,0.008012444608741337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,3072,0.3259893258412679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,2560,0.01699555582470364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,2048,0.007019555403126611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,2560,0.007613333563009898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,2560,0.3439022170172797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,2048,0.01586844523747762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,1536,0.005632888939645555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,2048,0.007240888973077138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,2048,0.3213040033976237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,1024,0.0046364445653226645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,1536,0.006647111227114995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,1536,0.3342595630221897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,1024,0.014132445057233175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,768,0.0042204442951414315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,1024,0.006591111007663939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,768,0.013448889056841532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,768,0.0063164445261160536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,768,0.3223653369479709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,512,0.012791111237472959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,1536,0.014549333188268872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,512,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,256,0.003612444632583194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,1024,0.3475457827250163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,512,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,256,0.2980995443132189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,128,0.0035111109415690103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,256,0.00564177789621883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,128,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,128,0.3055831061469184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3584,128,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,64,0.0033697777738173804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3584,32,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,64,0.011784888803958893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,256,0.012709333664841123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3584,32,0.012788444757461548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,65536,0.08949689070383708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,65536,0.2051457828945584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,16384,0.03100799851947361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,65536,0.350416898727417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,16384,0.05780533287260267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3584,512,0.32682755258348256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,16384,0.33903911378648544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,16384,0.020635555187861126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,12288,0.022129777404997084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,12288,0.044191999567879565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,12288,0.3395342297024197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,65536,0.0689893364906311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,10240,0.01885777711868286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,12288,0.016876444220542908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,10240,0.014942222171359591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,10240,0.0383840004603068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,8192,0.015399111641777886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,10240,0.3391955428653293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,8192,0.03267644511328803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,7168,0.013543999857372709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,8192,0.012703999876976013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,8192,0.34149599075317383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,7168,0.0298977792263031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,6144,0.012585777375433179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,7168,0.011369778050316704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,7168,0.32582932048373753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,6144,0.027446221974160936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,6144,0.011049778097205691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,5120,0.01108266661564509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,6144,0.3330390983157688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,5120,0.024791111548741657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,5120,0.010039110978444418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,4096,0.010088889135254754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,5120,0.33631377749972874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,4096,0.020607110526826646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,3584,0.008967999782827165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,4096,0.008722667064931657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,4096,0.3348524570465088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,3584,0.01925155520439148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,3584,0.008338666624493068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,3072,0.008268444074524773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,3584,0.33203376664055717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,3072,0.018227555685573153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,2560,0.007723555796676212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,3072,0.0079297779334916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,3072,0.35025421778361004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,2560,0.016918222109476726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,2048,0.006611555400821898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,2560,0.007693332930405934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,2560,0.32509422302246094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,2048,0.01549600064754486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,2048,0.007296889192528195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,1536,0.005224888937340842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,2048,0.33772889773050946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,1536,0.014183110660976835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,1024,0.004649777793222004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,1536,0.0069271110826068455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,1536,0.3364444573720296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,1024,0.01350044459104538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,768,0.0042168887125121225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,1024,0.006272888845867581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,1024,0.3372906578911676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,768,0.01311555587583118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,512,0.003867555409669876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,768,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,768,0.3025973372989231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,512,0.01275466630856196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,512,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,256,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,512,0.33565423223707413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,256,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,128,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,256,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,256,0.3218497700161404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,64,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,3072,128,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,128,0.012175111307038201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,3072,128,0.30087823337978786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,64,0.011776000261306763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,65536,0.07197778092490302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,3072,32,0.012623111406962076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,65536,0.20321155918969047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,16384,0.025507556067572698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,65536,0.33422311147054035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,16384,0.05578666925430298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,16384,0.02033866610791948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,16384,0.3199297851986355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,12288,0.022987556126382615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,3072,32,0.003555555724435382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,12288,0.016512889001104567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,10240,0.020062221421135795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,12288,0.33030133777194554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,65536,0.06750310791863336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,10240,0.03789510991838243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,8192,0.014997333288192749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,10240,0.3184693389468723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,10240,0.014145778285132514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,8192,0.03223377797338698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,7168,0.013421333498424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,8192,0.011789333489206104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,8192,0.3279413382212321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,7168,0.029690666331185236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,6144,0.01240355521440506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,7168,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,12288,0.04377333323160807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,6144,0.027095110880004034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,7168,0.4036106798383925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,5120,0.010815111299355825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,6144,0.33750756581624347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,6144,0.010374222364690568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,5120,0.0247013337082333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,5120,0.009718221922715506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,4096,0.009594666461149851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,5120,0.31801867485046387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,4096,0.019824888971116807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,3584,0.008722667064931657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,4096,0.008671110702885522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,4096,0.36634667714436847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,3584,0.018698667486508686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,3072,0.00797155582242542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,3584,0.008358221915033128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,3584,0.33759556876288516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,3072,0.01756177842617035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,2560,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,3072,0.00795999997191959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,3072,0.3504862255520291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,2560,0.016542222764756944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,2048,0.006066666709052191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,2560,0.32920977804395885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,2048,0.015262222952312894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,2048,0.007263110743628607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,2048,0.336533334520128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,1536,0.0053093334039052325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,1536,0.014186667071448432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,1536,0.3469288878970676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,1536,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,1024,0.013518222504191928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,2560,0.007377777662542131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,1024,0.006343110981914733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,768,0.00453599997692638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,1024,0.331986665725708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,768,0.013140444126394061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,768,0.3288035657670763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,768,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,1024,0.004600000050332812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,512,0.012797333300113678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,512,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,512,0.33340978622436523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,256,0.003601777884695265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,256,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,256,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,128,0.003236444460021125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,512,0.003919110943873723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,128,0.012097777591811286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2560,128,0.005922666854328579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,64,0.00328533330725299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,128,0.2953546577029758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2560,32,0.003343111111058129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,65536,0.06028978029886881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,32,0.012074666718641916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2560,256,0.3452942106458876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,65536,0.1651137802335951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,16384,0.022780444886949327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,65536,0.33189598719278973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,65536,0.06660444206661649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,16384,0.04610310991605123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,16384,0.019818666908476088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,12288,0.020474667350451153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,12288,0.035994665490256414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2560,64,0.0118231111102634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,10240,0.016768000192112394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,12288,0.3888862133026123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,12288,0.015595555305480957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,10240,0.03204088740878635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,8192,0.013711999687883588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,10240,0.013771555489963956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,10240,0.35882311397128636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,8192,0.02751733362674713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,7168,0.012521777715947894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,8192,0.33568088213602704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,8192,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,16384,0.3261724313100179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,6144,0.011485333244005838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,7168,0.011048888994587792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,7168,0.3329351213243273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,6144,0.023045332895384893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,5120,0.010199999643696679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,6144,0.010098666780524785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,6144,0.34698934025234646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,5120,0.020053333706325956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,4096,0.00906311141120063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,5120,0.009383999639087254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,5120,0.3102000024583605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,4096,0.01753155555990007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,3584,0.008110221889283922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,4096,0.008660444782839881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,7168,0.025341333614455328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,4096,0.3381297853257921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,3072,0.0073315559162033936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,3584,0.00833155545923445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,3584,0.3212782277001275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,3072,0.016184000505341422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,2560,0.006777777853939269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,3072,0.007901333272457123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,3072,0.3124426735772027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,2560,0.015142222245534262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,2560,0.3204373253716363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,2560,0.007613333563009898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,2048,0.005605333381228977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,2048,0.014343111051453484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,3584,0.0165928883685006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,2048,0.330913782119751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,1536,0.004909333255555895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,2048,0.006993778049945831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,1536,0.013539555172125498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,1536,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,1024,0.004256000121434529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,1536,0.34263910187615293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,1024,0.013058666553762225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,768,0.0038853333228164246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,1024,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,1024,0.33388710021972656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,768,0.012749333348539142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,512,0.003742222156789568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,768,0.330168883005778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,512,0.01242844429281023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,512,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,512,0.30740266376071507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,256,0.0032177778581778207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,256,0.012139555480745105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,256,0.33244445588853627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,256,0.005593777944644292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,128,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,768,0.006206222292449739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,128,0.011701333026091257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,64,0.003245333416594399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,2048,128,0.2887386745876736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,2048,128,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,2048,32,0.00350222239891688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,65536,0.049343999889161855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,32,0.011686222420798408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,65536,0.1617128849029541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,65536,0.34913333257039386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,65536,0.06465688678953382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,16384,0.019753777318530612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,16384,0.04376266731156243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,16384,0.30662311447991264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,16384,0.01921599937809838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,12288,0.017280000779363845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,2048,64,0.011770666473441653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,12288,0.0353999998834398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,12288,0.015473778049151102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,10240,0.015232889188660515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,12288,0.31773334079318577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,8192,0.013632000320487551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,10240,0.031410667631361223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,10240,0.01309333327743742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,10240,0.32298045688205296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,8192,0.027063111464182537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,8192,0.01148444414138794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,8192,0.3215457863277859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,7168,0.024822221861945257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,6144,0.010946667028797997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,7168,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,7168,0.3227084477742513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,6144,0.021937777598698933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,5120,0.009530666801664565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,6144,0.010088000032636855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,6144,0.2977866596645779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,5120,0.01997244523631202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,4096,0.008045333127180735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,5120,0.009667555491129557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,5120,0.310278230243259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,4096,0.017522666189405654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,7168,0.012131555212868584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,3584,0.007112888826264276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,4096,0.008636444807052612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,4096,0.309933344523112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,3072,0.0069751110341813825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,3584,0.01661422186427646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,3584,0.008279999925030602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,3072,0.015511999527613321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,3072,0.00792711145348019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,2560,0.006314666734801398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,3072,0.31260355313618976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,2560,0.015132443772421943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,2048,0.0057911111248864075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,2560,0.007624000310897827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,2560,0.32031022177802193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,2048,0.014114666316244336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,1536,0.004923555586073133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,2048,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,2048,0.3126133282979329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,1536,0.013473778135246701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,1024,0.0042568888101312844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,1536,0.006650666809744305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,1536,0.32144443194071454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,1024,0.012770666844315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,3584,0.3377644485897488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,768,0.0042035554846127825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,1024,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,1024,0.31971467865837944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,512,0.0038862224254343244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,768,0.012514666550689273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,768,0.3240355650583903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,512,0.01239466667175293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,512,0.3492266601986355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,512,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,256,0.0033679999825027254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,256,0.012048000262843238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,256,0.005942222144868638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,256,0.3464471234215631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,128,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,128,0.011762667033407422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,768,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,64,0.0028391112056043413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1536,128,0.31924711333380806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1536,128,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1536,32,0.0029440000653266907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,65536,0.03557777735922072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,64,0.011537777880827585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1536,32,0.012406222522258759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,65536,0.15982045067681208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,16384,0.015858666764365304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,65536,0.32197509871588814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,16384,0.042882667647467725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,16384,0.32084088855319554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,12288,0.013289777768982781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,12288,0.034627556800842285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,65536,0.06297689014010958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,12288,0.3292773299747043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,10240,0.012248000337017907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,12288,0.01441333360142178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,10240,0.030332446098327637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,16384,0.017836444907718234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,10240,0.013078221844302284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,8192,0.013134222063753339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,10240,0.34960267278883195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,8192,0.02544266647762722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,8192,0.011417778001891242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,7168,0.011741333537631564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,8192,0.30899643898010254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,7168,0.023786667320463393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,6144,0.010447999669445885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,7168,0.011111111276679568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,7168,0.33273598882887095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,6144,0.02161866592036353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,5120,0.009388444324334463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,6144,0.010016000105275048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,6144,0.3218675454457601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,5120,0.019905777441130746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,5120,0.009655999640623728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,4096,0.007868444754017724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,5120,0.33820801311069065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,4096,0.01741333305835724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,3584,0.007154666715198093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,4096,0.008657777474986183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,4096,0.33786487579345703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,3584,0.016506666938463848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,3072,0.006577777779764599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,3584,0.00833333366447025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,3584,0.3268302281697591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,2560,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,3072,0.01581333412064446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,3072,0.007751999629868402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,2560,0.014487110906177096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,3072,0.33353779051038956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,2048,0.0053004444473319584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,2560,0.3164168993631999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,2560,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,2048,0.013890667094124688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,1536,0.004675555560323927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,2048,0.007267555428875818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,2048,0.3243244489034017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,1536,0.013412444127930535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,1024,0.004209777961174647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,1536,0.00664533343580034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,1024,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,768,0.003859555555714501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,1024,0.31848621368408203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,768,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,768,0.006213333457708359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,768,0.32559288872612846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,512,0.003584889074166616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,512,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,512,0.33340978622436523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,512,0.0058853332367208265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,256,0.0032097777972618737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,256,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,1024,0.006305777778228124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,128,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,256,0.3135911093817817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,128,0.011750222080283694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,128,0.30120621787177193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,128,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,64,0.002874666617976295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,1024,32,0.0029271110478374693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,64,0.011742221812407175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,65536,0.02915999955601162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,1024,32,0.011772444678677453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,65536,0.15908532672458225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,1024,1536,0.41212444835238987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,65536,0.33284444279140896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,16384,0.01350311107105679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,1024,256,0.0057591112951437635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,65536,0.06027911106745402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,16384,0.04311911265055338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,12288,0.01185955521133211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,16384,0.01749955614407857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,16384,0.32294045554267037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,12288,0.034267554680506386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,12288,0.014430221584108142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,10240,0.010741333166758219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,12288,0.32019821802775067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,10240,0.030459556314680312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,8192,0.00961511085430781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,10240,0.013164444102181328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,10240,0.33936354849073624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,8192,0.025032000409232244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,8192,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,7168,0.009034666750166152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,8192,0.32625333468119305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,7168,0.023223999473783705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,7168,0.011032000184059143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,6144,0.00870666652917862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,7168,0.352034674750434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,6144,0.021685333715544805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,5120,0.00903200027015474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,6144,0.010042666561073726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,6144,0.34092089864942765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,5120,0.019966221517986722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,4096,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,5120,0.009696000152164036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,5120,0.34082667032877606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,4096,0.017268444101015728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,3584,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,4096,0.008671999805503422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,4096,0.34248799747890896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,3584,0.016520000166363187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,3072,0.006573333508438534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,3584,0.008328888979223039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,3584,0.3411875565846761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,3072,0.015855110353893705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,2560,0.005988444305128521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,3072,0.007732444339328342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,3072,0.3309111065334744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,2560,0.015201777219772339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,2048,0.0052791109515561005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,2560,0.007626666790909237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,2560,0.32511377334594727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,2048,0.01421511173248291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,2048,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,2048,0.3303768899705675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,1536,0.013125333521101209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,1536,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,1536,0.33064354790581596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,1024,0.00425866660144594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,1024,0.012757333616415659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,1024,0.3338017728593614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,1024,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,768,0.003915555361244413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,1536,0.004919111314747068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,768,0.01240888900227017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,512,0.003583111282851961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,768,0.32954133881462944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,768,0.006110222389300664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,512,0.012137778103351593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,256,0.0035422220826148987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,512,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,512,0.3240186638302273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,256,0.011690666278203329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,128,0.0030746666921509635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,256,0.005669333454635408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,256,0.314080900616116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,128,0.011729777687125735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,64,0.0031866667171319327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,768,32,0.0031884445084465873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,768,128,0.2947804398006863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,768,128,0.005790222022268508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,64,0.01147555559873581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,65536,0.024849777420361836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,768,32,0.011749332977665795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,16384,0.01145155562294854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,65536,0.0600186652607388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,65536,0.15701244937049016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,16384,0.041837334632873535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,65536,0.33597421646118164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,12288,0.009727111293209923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,16384,0.017109332813156974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,16384,0.3252462281121148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,10240,0.0118222220076455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,12288,0.32131555345323354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,12288,0.014155555102560254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,10240,0.029501332177056208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,8192,0.01036622209681405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,10240,0.013122667041089801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,10240,0.317159997092353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,8192,0.02518400053183238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,7168,0.009696889254781935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,8192,0.011674666570292579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,8192,0.37067021263970273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,7168,0.023080888721677992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,6144,0.008968888885445064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,7168,0.010688888529936472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,7168,0.3337199952867296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,12288,0.03416622347301907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,6144,0.021577777134047613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,5120,0.00831733312871721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,6144,0.0101742222905159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,6144,0.2795182334052192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,5120,0.01998133295112186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,4096,0.007675555845101674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,5120,0.009364444348547194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,5120,0.3206435574425591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,4096,0.017232000827789307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,3584,0.006969778074158563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,4096,0.008638222184446123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,3584,0.01651733285850949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,3584,0.008268444074524773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,3584,0.3420355584886339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,3072,0.006620444357395172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,3072,0.015565334094895257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,3072,0.3116702238718669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,2560,0.005980444451173146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,3072,0.007958222594526079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,2560,0.014783110883500842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,4096,0.30133067237006295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,2560,0.33492265807257754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,2560,0.007339555356237624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,2048,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,2048,0.3335075643327501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,2048,0.007314666277832455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,1536,0.004872888740566042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,1536,0.013179555535316467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,1536,0.3264382150438097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,1024,0.0042151109211974675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,2048,0.005224888937340842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,1024,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,1024,0.006344888773229387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,768,0.0036968886852264404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,1024,0.3088844352298313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,768,0.012373333175977072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,768,0.3180604510837131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,768,0.006247111078765657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,512,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,1536,0.006676444576846228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,512,0.012158222496509552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,256,0.0032311110860771606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,512,0.005956444475385878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,512,0.31188978089226616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,256,0.011810666984981961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,128,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,256,0.005584888988071018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,256,0.32010311550564235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,128,0.011786667009194693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,64,0.002962666667169995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,512,128,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,512,128,0.2968195544348823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,512,32,0.0029164445069101122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,65536,0.016082665986484952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,64,0.011726222104496427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,512,32,0.011408888631396823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,16384,0.010468444062603844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,65536,0.15732444657219782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,65536,0.053842667076322764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,65536,0.3301768832736545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,12288,0.009011555876996782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,16384,0.01718222267097897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,16384,0.04190311166975233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,16384,0.3307022253672282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,12288,0.032877332634396024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,10240,0.00904266701804267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,12288,0.014482667048772177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,12288,0.3329377704196506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,8192,0.008013333711359236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,10240,0.013200889031092325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,8192,0.024762666887707178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,8192,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,7168,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,8192,0.40276087654961484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,7168,0.022839110758569505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,7168,0.010671111444632212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,7168,0.3662613232930501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,6144,0.0074720001882976955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,10240,0.02914488977856106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,6144,0.02167555524243249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,10240,0.32891644371880424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,6144,0.010073777702119615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,6144,0.3380924330817328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,5120,0.01963911122745938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,4096,0.007655111451943715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,5120,0.009321778184837764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,5120,0.32966221703423393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,4096,0.017249777913093567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,3584,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,4096,0.00867377801073922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,4096,0.3224986659155952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,3584,0.01646933290693495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,3072,0.006321777900060018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,3584,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,3584,0.32318133778042263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,3072,0.015525332755512662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,3072,0.007647111184067196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,5120,0.008001777860853408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,3072,0.3280213408999973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,2560,0.014763555592960782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,2560,0.30406310823228627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,2560,0.007334222396214803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,2048,0.013844444519943662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,2048,0.3449591000874837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,2560,0.005649777750174205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,1536,0.004579555657174853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,2048,0.007207999626795451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,1536,0.01314133322901196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,2048,0.005236444373925527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,1024,0.0038782221575578055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,1536,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,1024,0.012736000120639801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,1024,0.35707643296983504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,1024,0.0063235556913746735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,768,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,768,0.012420444852775998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,768,0.006301333506902059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,512,0.0032275555034478507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,1536,0.33995556831359863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,512,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,512,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,256,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,512,0.32830844985114205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,256,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,256,0.005904888941182031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,128,0.0031635556370019913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,768,0.34249067306518555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,128,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,256,128,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,128,0.3087564574347602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,64,0.0028355556229750314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,256,32,0.0028400001012616684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,64,0.011405333048767514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,256,32,0.011391110718250275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,65536,0.014840889308187695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,256,256,0.3201253414154053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,65536,0.1527573267618815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,16384,0.009453333086437648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,65536,0.3460106584760878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,65536,0.05098933312628004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,16384,0.04018488857481215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,12288,0.008292444050312042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,16384,0.01722400055991279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,12288,0.032976001501083374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,10240,0.008971555365456475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,12288,0.01422044469250573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,12288,0.31982045703464085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,10240,0.02916533417171902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,8192,0.008309333688682979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,10240,0.012996444271670448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,8192,0.025015999873479206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,10240,0.4518008761935764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,7168,0.007709333466158972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,8192,0.31232889493306476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,8192,0.011388444238238864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,7168,0.010821333361996545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,7168,0.02271466619438595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,6144,0.007328888608349695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,7168,0.30524177021450466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,6144,0.02094488839308421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,6144,0.010045333041085137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,5120,0.007676444119877285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,6144,0.3284062279595269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,5120,0.0196115556690428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,4096,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,5120,0.3697520097096761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,4096,0.01721333298418257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,4096,0.008641777767075432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,3584,0.0069848886794514125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,16384,0.3057768874698215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,4096,0.352891551123725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,3584,0.01626311077011956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,3584,0.008015111088752747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,3072,0.006633777585294511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,3584,0.3036666711171468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,3072,0.015509333875444202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,3072,0.00793866647614373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,3072,0.3276551034715441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,2560,0.014496889379289416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,2560,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,2048,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,2048,0.013820444544156393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,2048,0.3464044464959039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,2560,0.00683111117945777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,1536,0.00424177779091729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,5120,0.009353777600659264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,2048,0.0069644442862934535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,1536,0.012744888663291931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,1024,0.0038542221817705366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,1536,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,1536,0.3008506562974718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,1024,0.01239111108912362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,768,0.0038720000949170855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,1024,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,1024,0.3288328912523058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,768,0.012116444607575735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,512,0.003346666693687439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,768,0.006210666563775804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,768,0.3139653205871582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,512,0.01181244436237547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,512,0.005769777629110549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,512,0.3327706654866536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,256,0.0028986665937635633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,256,0.011742221812407175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,2560,0.3134595553080241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,128,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,256,0.3331955538855659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,256,0.005906666732496685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,128,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,64,0.0028284444577164124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,256,128,128,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,128,32,0.0028826666788922418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,256,128,128,0.2999502287970649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,64,0.011411555939250521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,65536,0.013781333135233985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,16384,0.009386666946940953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,128,32,0.011437333292431302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,12288,0.008367111285527548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,65536,0.15608178244696722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,10240,0.007679999702506595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,12288,0.032243556446499295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,8192,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,10240,0.029247111744350855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,7168,0.006639110959238476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,8192,0.024792888098292883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,6144,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,7168,0.022982221510675218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,5120,0.006219555520349079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,6144,0.021328000558747187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,4096,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,5120,0.019630221856964957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,3584,0.006613333192136552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,4096,0.01723288827472263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,3072,0.0063279999627007385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,16384,0.04089511103100247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,2560,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,3072,0.015511110424995422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,2048,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,2560,0.014467555615637036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,1536,0.00453955555955569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,2048,0.013754666679435305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,1024,0.003912000192536248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,1536,0.012773333324326409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,768,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,1024,0.012453333371215396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,512,0.0035013332962989807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,768,0.012114666402339935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,256,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,512,0.012109333442317115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,128,0.002862222285734283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,3584,0.016216889023780823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,64,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,128,0.011712888876597086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,64,32,0.0028568889117903183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,64,0.011424000064531961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,65536,0.01349066694577535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,32,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,16384,0.008965333302815756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,65536,0.15512977706061468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,16384,0.03940000136693319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,12288,0.00832799987660514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,10240,0.007698666718271043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,12288,0.03231733375125461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,8192,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,64,256,0.01145155562294854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,7168,0.0070088886552386815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,8192,0.02478666603565216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,6144,0.006622222148709827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,5120,0.0064426664676931165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,6144,0.021327111456129286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,4096,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,5120,0.01977422171168857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,3584,0.006688000013430913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,4096,0.017217778497272067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,3072,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,3584,0.016572443975342643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,3072,0.015505777464972602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,2560,0.006695999867386288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,2048,0.005991999887757831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,2560,0.014491554763582019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,1536,0.004375111311674118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,7168,0.022729777627521094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,2048,0.013849777479966482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,1024,0.003920888735188378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,768,0.003532444437344869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,1536,0.013101333545313941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,512,0.003216888962520493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,1024,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,256,0.0031679999083280563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,768,0.012101333174440594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,128,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,512,0.012033777932325998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,64,0.002883555574549569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,256,0.011929777761300405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,256,32,32,0.0028515555378463534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,128,0.01108000013563368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,64,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,32,0.011781333221329583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,256,32,10240,0.028993778758578833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,16384,0.3877600034077962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,16384,0.2901342180040148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,12288,0.29224710994296604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,16384,0.13230844338734946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,12288,0.22409510612487793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,12288,0.0957111120223999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,16384,0.3606462213728163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,10240,0.21950666109720865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,10240,0.20297333929273817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,12288,0.347017765045166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,8192,0.17962222629123262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,10240,0.08299911022186279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,8192,0.15468177530500624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,10240,0.3470115661621094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,8192,0.07062311304940118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,7168,0.1584284438027276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,7168,0.13362310992346868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,8192,0.34764711062113446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,6144,0.1382924450768365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,7168,0.05887111028035482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,6144,0.11695822079976399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,7168,0.3449279997083876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,5120,0.11683200465308295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,6144,0.0530426667796241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,5120,0.09752711322572495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,6144,0.33612799644470215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,4096,0.09639288981755574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,5120,0.04514222343762716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,5120,0.34247644742329914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,4096,0.08470400174458821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,3584,0.08354577753278945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,4096,0.03832533293300205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,3584,0.07260533173878987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,4096,0.3346150981055365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,3584,0.0340302222304874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,3072,0.07190755340788099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,3584,0.3258506721920437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,2560,0.06198222107357449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,3072,0.06601866748597887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,3072,0.03036266565322876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,3072,0.32769155502319336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,2560,0.05664622121387058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,2048,0.05379822187953525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,2560,0.027186666925748188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,2048,0.04879022306866116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,2560,0.334166235393948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,1536,0.03884977764553494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,2048,0.023429332507981196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,1536,0.04145866632461548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,1024,0.027906666199366253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,1536,0.018383999665578205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,1536,0.33067822456359863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,1024,0.03394844465785556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,768,0.022690667046440974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,1024,0.3135582341088189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,768,0.028464890188641016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,768,0.013528889252079858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,512,0.016270221935378183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,2048,0.3243244489034017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,512,0.023994666006830003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,512,0.01239111108912362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,256,0.013521778086821238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,1024,0.014477334088749357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,256,0.02186044388347202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,256,0.011019555230935415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,256,0.29154666264851886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,768,0.31748710738288033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,128,0.020654223031467862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,128,0.3014124499426948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,65536,128,0.010992888775136737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,64,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,32,0.012155555188655853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,65536,512,0.30182843738132054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,32,0.02195022172398037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,65536,0.36814043256971574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,65536,128,0.011359111302428775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,65536,0.370344877243042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,16384,0.09813155730565389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,16384,0.1056257751252916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,65536,0.14268267154693604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,65536,64,0.020634666085243225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,12288,0.07397688759697808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,16384,0.04352444410324097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,12288,0.0801102254125807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,16384,0.288881778717041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,10240,0.06470488839679293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,12288,0.03238933285077413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,12288,0.31472799513075084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,10240,0.06774310933219062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,8192,0.05175288849406772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,65536,0.4617990917629666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,10240,0.028933333026038274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,10240,0.33394667837354874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,8192,0.02484888831774394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,8192,0.05918222003512912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,8192,0.2875902122921414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,7168,0.05276266733805338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,6144,0.04147555430730184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,7168,0.3310391108194987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,7168,0.02145066691769494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,6144,0.04367555512322319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,6144,0.019090667366981506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,6144,0.33535112275017637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,5120,0.034669333034091525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,5120,0.03896711270014445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,5120,0.3460471100277371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,5120,0.01690933273898231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,4096,0.02812355425622728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,4096,0.03472888800832961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,4096,0.34527378612094456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,3584,0.026168000366952684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,3584,0.031082666582531396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,7168,0.04675022098753187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,3584,0.01330844478474723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,3072,0.022482666704389784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,3584,0.3502151171366374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,3072,0.027779555983013574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,4096,0.014411555396185981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,3072,0.33735378583272296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,2560,0.019709333777427673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,2560,0.02459999918937683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,2560,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,2048,0.017635555730925668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,2048,0.02200800014866723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,2048,0.34830400678846574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,3072,0.012489777472284106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,1536,0.013045333325862885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,2048,0.010580444501505958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,1536,0.019631110959582858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,2560,0.35193689664204914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,1024,0.01016533292002148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,1536,0.008386666576067606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,1536,0.340923547744751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,1024,0.01757333344883389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,768,0.008735111190213097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,1024,0.00794933322403166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,1024,0.3357386589050293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,768,0.016177778442700703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,512,0.006945778098371293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,768,0.007292444507280986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,768,0.33483110533820254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,512,0.015097777048746744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,256,0.00489688871635331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,512,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,512,0.33167823155721027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,256,0.014175110393100314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,128,0.004247111164861255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,256,0.006591111007663939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,256,0.32828622394137913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,64,0.004926222066084544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,128,0.013755555782053204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,16384,128,0.006588444527652528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,16384,32,0.004565333326657613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,64,0.014135110709402295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,16384,32,0.013764444324705334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,16384,128,0.4371795654296875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,65536,0.2643582291073269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,16384,0.07169600327809651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,65536,0.28828001022338867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,65536,0.1355946726269192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,16384,0.08204178015391032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,12288,0.056185775332980685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,16384,0.040327999326917864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,16384,0.3358888891008165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,12288,0.06493244568506877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,10240,0.04756266540951199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,12288,0.03030311067899068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,10240,0.05457511213090685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,12288,0.3807066546546088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,8192,0.03947733177079095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,10240,0.026305778159035578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,10240,0.3378888765970866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,8192,0.04857244425349765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,7168,0.03517599900563558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,8192,0.02293955617480808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,8192,0.3354489008585612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,7168,0.04316799839337667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,6144,0.031405333015653826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,7168,0.01962311069170634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,7168,0.32731199264526367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,6144,0.03602666656176249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,5120,0.026759111218982275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,65536,0.44805065790812176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,6144,0.018107554978794523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,6144,0.32989244990878636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,4096,0.02241866621706221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,5120,0.032293332947625056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,5120,0.016206221448050607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,4096,0.027671999401516382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,3584,0.019537778364287484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,4096,0.014167111780908374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,4096,0.3194551202985975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,3584,0.025019556283950806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,3072,0.016872889465755887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,3584,0.013163555827405719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,3584,0.32757777637905544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,3072,0.02295644415749444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,2560,0.014898666077189975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,3072,0.012373333175977072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,3072,0.3212089008755154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,2560,0.021320889393488567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,2560,0.3059599929385715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,2560,0.010899555351999072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,5120,0.31941954294840497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,2048,0.013159111142158508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,2048,0.019658666517999437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,2048,0.009726222190592024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,1536,0.01074488874938753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,2048,0.31132178836398655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,1536,0.01796800063716041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,1024,0.008567999634477828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,1536,0.3440373208787706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,1536,0.008078222473462423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,1024,0.01588355501492818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,768,0.007602666815121968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,1024,0.00740799970097012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,1024,0.3271964391072591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,768,0.014877332581414117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,512,0.005752888818581899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,768,0.007268444531493717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,768,0.33379199769761825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,512,0.014153778553009033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,256,0.004598222259018156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,512,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,512,0.3371475537618001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,256,0.013540444274743399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,128,0.004239999999602635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,256,0.006629333313968446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,128,0.013415111435784234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,12288,128,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,128,0.31274133258395725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,64,0.0038559999730851916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,12288,32,0.0038924444880750445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,64,0.013408888545301227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,12288,32,0.015212444795502557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,65536,0.27280712127685547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,12288,256,0.3290453222062853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,65536,0.12875289387173122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,16384,0.07641600237952338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,16384,0.07926577991909452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,65536,0.44678576787312824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,65536,0.27556800842285156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,12288,0.055778668986426465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,16384,0.038451555702421404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,12288,0.0627431141005622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,10240,0.04907288816240099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,12288,0.029483556747436523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,12288,0.3265475432078044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,10240,0.053575111760033496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,10240,0.02533600065443251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,8192,0.03986577855216132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,10240,0.3461475637223985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,8192,0.04698044392797682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,7168,0.03516622384389242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,8192,0.022103110949198406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,16384,0.33162932925754124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,8192,0.3013919989267985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,6144,0.03093866507212321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,7168,0.04114222195413377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,7168,0.01924088928434584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,6144,0.03492533498340183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,7168,0.3368106683095296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,5120,0.02701866626739502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,6144,0.017439110411538016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,6144,0.3240453402201335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,5120,0.030615111192067463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,4096,0.022520000735918682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,5120,0.015834665960735746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,5120,0.3199911117553711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,3584,0.019879110985332064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,4096,0.013843555417325763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,4096,0.33224354849921334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,3584,0.0247715562582016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,3072,0.018037334084510803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,3584,0.01274311128589842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,3584,0.3407333427005344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,3072,0.02294844388961792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,2560,0.01568711135122511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,3072,0.011591110792424945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,3072,0.31147999233669704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,2560,0.021074665917290583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,2560,0.01038044442733129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,2560,0.32386933432685006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,2048,0.019114666514926486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,4096,0.02678666677739885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,1536,0.010332444475756751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,2048,0.008676444490750631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,2048,0.295106675889757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,1536,0.017343999611006845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,1024,0.00833777752187517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,1536,0.007902222375075022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,1536,0.33307909965515137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,1024,0.015178667174445258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,768,0.007035555938879649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,1024,0.007442666424645319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,1024,0.3219288984934489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,768,0.014491554763582019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,2048,0.013749333719412485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,512,0.005643555687533484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,768,0.006989333364698622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,512,0.013805333111021252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,768,0.3350737889607747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,256,0.004864888886610667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,512,0.33193421363830566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,512,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,256,0.01347733371787601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,128,0.0042160000238153665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,256,0.006598222172922558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,256,0.32601332664489746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,128,0.012809777425395118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,64,0.003866666720973121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,10240,128,0.29726844363742405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,10240,128,0.006633777585294511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,10240,32,0.0038888889054457345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,64,0.013178666432698568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,10240,32,0.01347555551264021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,65536,0.1841599941253662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,65536,0.23393689261542427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,16384,0.049738665421803795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,65536,0.08222488562266032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,16384,0.06918222374386258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,16384,0.325638214747111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,65536,0.34975022739834255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,12288,0.04085422224468655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,12288,0.05456977751519945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,12288,0.3153822157118055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,10240,0.03516088922818502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,12288,0.019658666517999437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,10240,0.04708355665206909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,10240,0.3106951183742947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,10240,0.017270222306251526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,8192,0.027671110298898485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,16384,0.02545333405335744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,8192,0.04047288828425937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,8192,0.015606222881211175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,7168,0.025352888637118872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,8192,0.31607643763224286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,7168,0.03492177857293023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,6144,0.022706665926509436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,7168,0.013857777747843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,7168,0.32823022206624347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,6144,0.02975644336806403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,5120,0.019172444939613342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,6144,0.32308533456590444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,5120,0.02671555512481266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,5120,0.33092355728149414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,5120,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,4096,0.017695999807781644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,4096,0.024173332585228816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,4096,0.32928888003031415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,4096,0.01036177741156684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,3584,0.01422755585776435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,6144,0.012996444271670448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,3584,0.02204088866710663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,3072,0.012730666332774691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,3584,0.3220462269253201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,3072,0.02088977727625105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,3584,0.009680888719028896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,2560,0.011427555647161273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,3072,0.00902844468752543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,3072,0.31019555197821724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,2048,0.010425777898894416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,2560,0.019288889235920377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,2560,0.008237332933478886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,2560,0.33043111695183647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,2048,0.017927999297777813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,1536,0.008345777789751688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,2048,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,1536,0.01612444387541877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,1024,0.006586666736337874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,1536,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,2048,0.40756977929009336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,1536,0.32706133524576825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,768,0.005230222311284807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,1024,0.01443733274936676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,1024,0.0063662222690052455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,1024,0.31595200962490505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,768,0.014088888963063559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,512,0.004600888739029567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,768,0.00637511122557852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,768,0.3166648811764187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,512,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,256,0.0038648889296584656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,512,0.005998222364319696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,512,0.3133688767751058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,256,0.013049778011110095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,128,0.003572444534964032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,256,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,256,0.3129928906758626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,128,0.01253600004646513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,64,0.0035751110149754416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,8192,32,0.0039022221333450745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,8192,128,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,8192,128,0.2990506754981147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,64,0.012746666868527731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,65536,0.1719671090443929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,8192,32,0.014122666584120857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,16384,0.048249777820375234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,65536,0.22377332051595053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,16384,0.06768266359965007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,16384,0.359482659233941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,65536,0.3451208803388808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,12288,0.03804711169666714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,16384,0.024677332904603746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,12288,0.05371822251213921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,12288,0.3574879964192708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,12288,0.018608889646000333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,10240,0.03243288728925917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,65536,0.07822933461931017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,10240,0.04582755433188545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,10240,0.016618667377365958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,8192,0.026574222577942744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,10240,0.32934488190544975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,8192,0.03815822137726678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,7168,0.02332177758216858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,8192,0.015116444892353482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,8192,0.34114755524529355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,7168,0.03350311186578538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,6144,0.021595555875036452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,7168,0.01350044459104538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,7168,0.3426160017649333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,6144,0.029020445214377508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,6144,0.012675555215941535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,5120,0.017482666505707633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,6144,0.3366240130530463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,5120,0.026192888617515564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,5120,0.011408888631396823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,4096,0.015531554818153381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,5120,0.3301733334859212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,4096,0.024047111471494038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,3584,0.013575110998418597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,4096,0.009859555297427708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,4096,0.3393511242336697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,3584,0.02195377813445197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,3072,0.0123857781291008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,3584,0.009257777697510189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,3584,0.3225422170427111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,3072,0.020635555187861126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,2560,0.010891555911964841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,3072,0.008282666405042013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,3072,0.3378257751464844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,2560,0.01904177831278907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,2048,0.010019555687904358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,2560,0.008216000265545314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,2560,0.32575999365912545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,2048,0.01720266706413693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,1536,0.007937777373525832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,2048,0.007285333342022366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,2048,0.3329271210564507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,1536,0.015805333852767944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,1024,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,1536,0.006966222491529252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,1536,0.3189511034223768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,1024,0.014431110686726041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,768,0.0058471109304163195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,1024,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,1024,0.32344267103407115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,768,0.013808888693650564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,512,0.004899555610285865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,768,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,768,0.3208186626434326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,512,0.013467555244763693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,256,0.0041786668201287585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,512,0.00591822216908137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,512,0.32475021150377065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,128,0.0038648889296584656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,256,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,256,0.01294133315483729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,256,0.330222209294637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,64,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,7168,128,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,7168,128,0.30004000663757324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,7168,32,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,64,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,32,0.013081777426931592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,65536,0.17551022105746797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,65536,0.19197689162360299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,16384,0.047409776184293956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,65536,0.07599021991093953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,16384,0.0580293337504069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,7168,128,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,65536,0.31370666291978627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,12288,0.036905778778923884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,16384,0.3773653242323134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,12288,0.04580266608132256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,10240,0.032263110081354775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,12288,0.018072000808186002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,12288,0.31527911292182076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,10240,0.03823466764556037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,10240,0.016247999336984422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,8192,0.027253333065244887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,10240,0.34485332171122235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,8192,0.032927112446890935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,7168,0.023847111397319373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,8192,0.014710222681363424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,8192,0.325363556543986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,16384,0.022788445154825848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,7168,0.029343999094433252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,6144,0.02091644373204973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,7168,0.01315822203954061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,7168,0.34759823481241864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,6144,0.025637333591779072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,5120,0.017795556121402316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,6144,0.29994577831692165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,6144,0.012261333564917246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,5120,0.023221333821614582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,4096,0.015466666883892484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,5120,0.010915555887752108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,5120,0.3000444571177165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,4096,0.021361778179804485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,3584,0.013292444248994192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,4096,0.34316089418199325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,3584,0.019679110911157396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,3584,0.00869511150651508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,3584,0.33854489856296116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,3072,0.011872889267073737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,3072,0.018952889574898612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,3072,0.34412444962395566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,3072,0.008292444050312042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,2560,0.010974222587214576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,4096,0.009693333672152625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,2560,0.017705778280893963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,2048,0.009682666924264696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,2560,0.3398844401041667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,2048,0.015978667471143935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,2048,0.33342843585544163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,2048,0.007391999993059371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,1536,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,1536,0.01462933255566491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,1536,0.32355109850565594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,1536,0.0069324444565508105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,1024,0.005873777800136142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,1024,0.013767999907334646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,2560,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,768,0.005231110999981563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,1024,0.33319467968410915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,1024,0.006280000011126201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,768,0.013439999686347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,512,0.004247111164861255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,768,0.006245333287451003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,768,0.32235466109381783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,512,0.012826667063766055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,256,0.003847111016511917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,512,0.005954666684071223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,512,0.33826221360100645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,256,0.012506666282812754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,128,0.00388355553150177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,256,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,256,0.3289742204878065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,128,0.012118221984969245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,64,0.0032835555159383347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,6144,128,0.00589155571328269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,6144,128,0.2958204481336805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,6144,32,0.0035084444615576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,64,0.012252444194422828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,6144,32,0.013064000341627332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,65536,0.1427946620517307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,65536,0.18753333886464438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,16384,0.045138667027155556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,65536,0.07324800226423475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,16384,0.0562266641192966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,12288,0.03542044427659776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,16384,0.02174844510025448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,16384,0.3317786587609185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,12288,0.04378133349948459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,12288,0.01755466726091173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,10240,0.03083289000723097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,10240,0.03688888748486837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,12288,0.4528906610276964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,8192,0.025694222913848028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,10240,0.3286302354600695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,10240,0.015948444604873657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,8192,0.03159644537501865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,8192,0.014144000079896716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,7168,0.022275555464956496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,8192,0.3211013211144341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,65536,0.3276746537950304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,7168,0.02888800038231744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,6144,0.020036444067955017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,7168,0.012854222622182635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,6144,0.025420443879233465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,7168,0.361861334906684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,6144,0.011396444506115384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,5120,0.016697777642144095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,6144,0.32107732031080455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,4096,0.014056000444624158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,5120,0.023031999667485554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,5120,0.010392888552612728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,5120,0.30570843484666615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,4096,0.02160088883505927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,4096,0.009018667042255402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,4096,0.3407306671142578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,3584,0.01953866746690538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,3072,0.01146133326821857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,3584,0.33739378717210555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,3072,0.018392889036072623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,3072,0.00795377790927887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,2560,0.010218666659461128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,3072,0.3341226577758789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,2560,0.01677688956260681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,2560,0.32706043455335826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,2560,0.007715555528799693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,2048,0.00923822240697013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,3584,0.008352888955010308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,2048,0.01553066737122006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,2048,0.007357333269384172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,1536,0.007268444531493717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,1536,0.014426666829321118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,3584,0.012792000340090858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,1536,0.006937777830494775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,1536,0.3408168951670329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,1024,0.00554044461912579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,1024,0.013506666653686099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,1024,0.32995110087924534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,1024,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,768,0.00498577786816491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,2048,0.31967822710673016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,768,0.013190222283204397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,768,0.006243555496136348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,512,0.004869333157936732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,768,0.336899545457628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,512,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,256,0.0038559999730851916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,512,0.005985777825117111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,512,0.33751466539171004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,256,0.012173333101802401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,128,0.0035217776894569397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,256,0.005882666756709416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,256,0.32813334465026855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,128,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,64,0.0032248888164758682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,5120,128,0.00589155571328269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,5120,32,0.0032248888164758682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,5120,128,0.31396712197197807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,64,0.012140444583363004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,65536,0.10541688733630711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,5120,32,0.0123848890264829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,65536,0.1772915522257487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,65536,0.07031822204589844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,16384,0.0346942212846544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,16384,0.053416000472174756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,65536,0.34253599908616805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,12288,0.02892711096339756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,16384,0.3277440071105957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,16384,0.021129777034123737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,12288,0.040892445378833346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,12288,0.32345422108968097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,10240,0.024693333440356787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,10240,0.035190221336152815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,10240,0.33968355920579696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,10240,0.01534844438234965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,8192,0.020627554919984605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,8192,0.03092977735731337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,7168,0.01904800037542979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,8192,0.3258231216006809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,8192,0.01313688854376475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,12288,0.01717155509524875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,7168,0.028441776831944782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,6144,0.01662133302953508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,7168,0.3223777876959907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,6144,0.02495199938615163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,6144,0.011090666883521609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,5120,0.014498665928840637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,6144,0.30541600121392143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,5120,0.022456000248591106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,5120,0.010036444498433007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,4096,0.012672888735930124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,5120,0.4043262269761827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,4096,0.34217156304253477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,4096,0.008665777742862701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,3584,0.009670221971140968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,7168,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,3584,0.019215111931165058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,3072,0.008752000000741746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,3584,0.00831822223133511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,3584,0.3291422261132134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,3072,0.01752000053723653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,3072,0.008068444828192392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,2560,0.008009778128729926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,3072,0.32143643167283803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,2560,0.016146666473812528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,4096,0.020789333515697055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,2048,0.0069448889957533936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,2560,0.007616889145639207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,2560,0.34968267546759707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,1536,0.005437333136796951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,2048,0.015118221441904703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,2048,0.007163555257850223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,2048,0.33332978354560006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,1024,0.004651555584536659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,1536,0.32939378420511883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,1024,0.01348622226052814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,1024,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,768,0.0042160000238153665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,1024,0.33082487848069936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,768,0.013175999952687157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,768,0.3250524467892117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,768,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,512,0.0038195554580953387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,1536,0.014436443646748861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,1536,0.006696888970004187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,512,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,256,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,512,0.3109084500206841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,512,0.005934222290913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,256,0.012449777788586087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,128,0.0035537779331207275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,256,0.005640000104904175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,256,0.31876177257961696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,128,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,64,0.003198222153716617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,4096,128,0.00564533347884814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,4096,128,0.297269344329834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,4096,32,0.003522666792074839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,64,0.012151111331250934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,65536,0.1261404487821791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,4096,32,0.011800000237094032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,65536,0.1785128911336263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,65536,0.06901244322458903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,16384,0.03892622060245938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,16384,0.05284355415238274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,65536,0.3418666786617703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,12288,0.029193778832753498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,16384,0.0207368897067176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,12288,0.04092266824510362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,12288,0.30075110329522026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,12288,0.017024889588356018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,10240,0.02230311102337307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,10240,0.03493511014514499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,10240,0.3048657841152615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,8192,0.018769777483410306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,10240,0.015395555231306287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,16384,0.34791111946105957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,8192,0.2992417812347412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,7168,0.018353778454992507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,8192,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,7168,0.028221332364612158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,7168,0.3366986645592584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,7168,0.011959999799728394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,6144,0.016182222300105624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,6144,0.025015999873479206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,6144,0.33301599820454914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,6144,0.010390222072601318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,8192,0.030774222479926214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,5120,0.014207111464606391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,5120,0.010191111101044549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,4096,0.012198222180207571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,5120,0.2998888757493761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,4096,0.02027466727627648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,4096,0.30488978491889107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,4096,0.008810666700204214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,3584,0.010191111101044549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,3584,0.01830400029818217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,3072,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,3584,0.2995679908328586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,3584,0.008480889101823172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,3072,0.016898666818936665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,2560,0.007431999676757389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,5120,0.0224906669722663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,3072,0.2929111056857639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,3072,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,2560,0.015871110889646742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,2048,0.0069777775141927935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,2560,0.007601777712504069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,2560,0.30546487702263725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,2048,0.015218666858143277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,2048,0.007122666471534305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,1536,0.014161777165200977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,2048,0.42024532953898114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,1024,0.004622222234805425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,1536,0.31414222717285156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,1536,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,1024,0.013632000320487551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,1024,0.006544000158707301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,768,0.004902222090297275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,1024,0.3132559988233778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,768,0.013091555900043912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,512,0.004250666747490565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,768,0.006166222194830577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,768,0.3669279946221246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,1536,0.00553955551650789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,512,0.012831110921170978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,256,0.003552888830502828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,512,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,512,0.31812887721591526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,256,0.012288000020715924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,128,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,256,0.005590222362014983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,256,0.31259200308057994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,128,0.012063999970753988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,64,0.0032462223122517266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3584,32,0.0032071111102898917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3584,128,0.005906666732496685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3584,128,0.2967528767055935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,64,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,65536,0.09188089105818008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3584,32,0.012097777591811286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,16384,0.030086222622129653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,65536,0.13837422264946833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,65536,0.06826666990915935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,16384,0.0418488879998525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,65536,0.33986843956841367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,16384,0.02068622244728936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,12288,0.021197333931922913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,12288,0.032524443335003324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,12288,0.34598933325873477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,10240,0.018027555611398485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,10240,0.028909332222408716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,16384,0.33615022235446507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,10240,0.3382568889194065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,8192,0.015068444940778943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,10240,0.014797333214018079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,8192,0.02631644407908122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,12288,0.016881777180565726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,8192,0.32504354582892525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,7168,0.013479111095269522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,7168,0.02397511071628994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,7168,0.34336445066663956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,7168,0.01181688904762268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,6144,0.01260444439119763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,6144,0.020940444535679285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,6144,0.33147554927402073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,5120,0.010451555252075195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,8192,0.012735111018021902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,5120,0.019630221856964957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,5120,0.009967111051082611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,4096,0.009762666291660732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,5120,0.3541395664215088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,4096,0.01683644453684489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,4096,0.00869688888390859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,4096,0.3363582293192546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,3584,0.008664888640244802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,6144,0.01034311122364468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,3584,0.016535111599498324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,3072,0.007994666695594788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,3584,0.00798044436507755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,3584,0.3364391061994765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,3072,0.01517777807182736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,2560,0.007220444579919179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,3072,0.008052444292439355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,3072,0.31008889940049916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,2560,0.014509333504570855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,2048,0.00628977765639623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,2560,0.007623111208279927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,2560,0.32540443208482533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,2048,0.013927110367351107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,1536,0.00536088893810908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,2048,0.007258666886223688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,2048,0.30816978878445095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,1024,0.00453333349691497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,1536,0.0069226668112807805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,1024,0.012810666528013019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,1536,0.47215912077162003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,1024,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,1024,0.3245297802819146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,768,0.004277333203289243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,768,0.01240533341964086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,512,0.0038977778620190094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,768,0.3224373393588596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,512,0.01240533341964086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,512,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,512,0.2896568775177002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,256,0.0033866665843460294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,256,0.012106666962305704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,256,0.3287893401251899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,256,0.005559111220969095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,128,0.0032213332338465583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,128,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,768,0.00630044440428416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,64,0.0032542221662071017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,3072,128,0.2991742293039958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,3072,128,0.0058648888435628675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,3072,32,0.003154666680428717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,65536,0.09348533550898235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,64,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,32,0.011808888779746162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,3072,1536,0.013467555244763693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,16384,0.031028442912631567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,65536,0.1368737750583225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,65536,0.06697689162360297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,16384,0.04052444299062093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,12288,0.023779556155204773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,16384,0.020226667324701946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,65536,0.39189423455132383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,16384,0.3232204384273953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,10240,0.01683644453684489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,12288,0.03191733360290527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,12288,0.016238222519556682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,12288,0.331425772772895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,10240,0.02823466724819607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,10240,0.014190221826235453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,10240,0.3325128820207384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,8192,0.025660443637106154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,7168,0.012479999827014076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,8192,0.01220444424284829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,8192,0.30308355225457084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,7168,0.02367288867632548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,6144,0.011545778148704104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,7168,0.0110880004035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,7168,0.33555110295613605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,6144,0.02065955599149068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,6144,0.3239440123240153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,6144,0.010027555955780877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,5120,0.010263111028406356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,8192,0.013922666509946188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,5120,0.018862222631772358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,4096,0.009744889206356471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,5120,0.009658666948477427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,5120,0.3238817850748698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,4096,0.01651822196112739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,4096,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,3584,0.008430221842394935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,4096,0.31994578573438853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,3584,0.015984889533784654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,3072,0.008096888661384583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,3584,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,3584,0.3312151167127821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,3072,0.0151573336786694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,2560,0.007281777759393056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,3072,0.007673777639865875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,3072,0.33014488220214844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,2560,0.014440889159838358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,2560,0.007607999775144789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,2560,0.33536354700724286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,2048,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,2048,0.32022666931152344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,2048,0.007285333342022366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,1536,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,1536,0.01350400017367469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,1536,0.32569776640997994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,1536,0.006736889067623351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,1024,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,2048,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,1024,0.3180239995320638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,768,0.004259555704063839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,768,0.012411555482281579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,768,0.32795466317070854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,768,0.005993777679072485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,512,0.003924444317817688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,1024,0.012855999999576144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,512,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,1024,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,256,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,512,0.3127857844034831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,256,0.01204711116022534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,256,0.0058346668051348785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,256,0.30654221110873753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,128,0.003212444484233856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,128,0.011415999796655444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2560,128,0.29626843664381236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,128,0.005606222069925732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,64,0.003330666571855545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2560,32,0.0032044444233179092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,64,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,65536,0.06415910853279962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2560,32,0.012132444315486483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2560,512,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,16384,0.02271466619438595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,65536,0.06548888815773858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,65536,0.334370666080051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,16384,0.03932888971434699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,12288,0.020052444603708055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,16384,0.019952888290087383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,12288,0.031876444816589355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,16384,0.39362666341993546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,10240,0.016562667157914903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,12288,0.0162444445821974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,12288,0.3177448908487956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,10240,0.02777422302299076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,8192,0.013565333353148567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,10240,0.014121777481502958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,10240,0.309824890560574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,65536,0.13499732812245688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,7168,0.012499555945396423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,8192,0.024919999970330134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,8192,0.011710222396585675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,8192,0.32892711957295734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,7168,0.023335999912685815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,6144,0.011399110986126794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,7168,0.010690666735172272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,7168,0.3449466758304172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,6144,0.020536889632542927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,6144,0.010068444742096795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,5120,0.009751111268997192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,6144,0.3191724353366428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,4096,0.00866844422287411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,5120,0.009680888719028896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,5120,0.3057404359181722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,4096,0.016233777006467182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,3584,0.00776800016562144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,4096,0.008634666601816813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,4096,0.34677600860595703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,3584,0.0158906661801868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,3072,0.007155555817815993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,3584,0.008395555946562026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,3584,0.30492178599039715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,3072,0.014911111858155994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,2560,0.006281777802440856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,3072,0.007643555601437886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,3072,0.28072799576653373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,2560,0.014475555883513557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,5120,0.017863111363516915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,2048,0.005516444643338521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,2560,0.3212249014112684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,2048,0.013767999907334646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,2048,0.006992888947327931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,2048,0.3162142170800103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,1536,0.005063999858167437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,1536,0.01309333327743742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,1536,0.33297599686516655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,1536,0.006601777755551868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,1024,0.004625777817434735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,1024,0.012422222230169507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,2560,0.0075795559419526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,768,0.003903111235962974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,1024,0.31857866711086696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,1024,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,768,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,768,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,512,0.003842666745185852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,768,0.3214968840281169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,512,0.012128888732857175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,256,0.0035413333939181436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,512,0.005910222315125995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,512,0.3287431134117974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,256,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,128,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,256,0.005665777872006099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,256,0.3200613392723931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,128,0.011400889191362592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,64,0.0028648889727062653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,2048,128,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,2048,32,0.0031662223239739737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,2048,128,0.3088231086730957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,64,0.011657777759763928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,2048,32,0.012072888513406118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,65536,0.13204977247450086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,65536,0.30433866712782115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,65536,0.06432000133726332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,16384,0.037823110818862915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,65536,0.04954666561550564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,16384,0.3424853218926324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,12288,0.015562666787041558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,16384,0.019108444452285767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,12288,0.03124888737996419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,16384,0.018046221799320646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,12288,0.33387290106879336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,12288,0.014488000008794995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,10240,0.027450667487250432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,10240,0.013083555632167392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,8192,0.013576889203654395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,8192,0.02403822210099962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,10240,0.4862693150838216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,8192,0.33148354954189724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,7168,0.012183999849690331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,8192,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,7168,0.0221404449807273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,7168,0.010738666686746808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,6144,0.01108000013563368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,6144,0.019318222999572754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,10240,0.013890667094124688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,5120,0.00979911122057173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,6144,0.3197840054829915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,6144,0.010106667048401302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,5120,0.01788444485929277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,5120,0.009735999835862054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,4096,0.00832444429397583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,5120,0.2959102259741889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,4096,0.016176000237464905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,3584,0.007264888948864407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,4096,0.3246213330162896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,3584,0.015471999843915304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,3584,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,7168,0.41304532686869305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,3584,0.3175537851121691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,3072,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,3072,0.01478666729397244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,2560,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,3072,0.303513765335083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,2560,0.014156444205178155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,4096,0.008664000365469191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,2560,0.3064302338494195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,2560,0.007611555357774098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,2048,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,2048,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,2048,0.0070257774657673305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,1536,0.004903111192915174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,2048,0.33563823170132107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,1536,0.013155555559529198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,1024,0.004190222256713443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,1536,0.006895111252864202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,1536,0.31286488638983834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,3072,0.007907555335097844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,768,0.0038568890757030914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,1024,0.012702222499582501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,1024,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,768,0.012153777811262341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,512,0.003535110917356279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,768,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,768,0.3257173432244195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,512,0.012051555845472546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,512,0.005983110931184556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,256,0.003208000005947219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,256,0.01183555523554484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,256,0.3103591071234809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,1024,0.32848623063829213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,128,0.0032097777972618737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,256,0.005587555468082428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,128,0.011708444191349877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,512,0.3267342249552409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1536,128,0.29023999638027614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,64,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1536,128,0.005943111247486538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1536,32,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,65536,0.03758666581577725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,64,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1536,32,0.011763555308183035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,65536,0.12966044743855795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,65536,0.062183108594682485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,16384,0.014367999302016364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,65536,0.32712801297505695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,16384,0.03773155477311876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,12288,0.012906666431162091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,16384,0.017186666528383892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,16384,0.3373600112067328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,12288,0.03022222386466132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,12288,0.014456000592973499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,10240,0.011516444385051727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,12288,0.3215804364946154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,10240,0.027092445227834914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,10240,0.013076444466908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,8192,0.010848888920413123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,10240,0.3220115502675374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,8192,0.02353511088424259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,7168,0.010081777969996134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,8192,0.011417778001891242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,8192,0.32663199636671275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,7168,0.021905778182877436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,6144,0.01017955525053872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,7168,0.33481423060099286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,6144,0.018933332628673978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,6144,0.010101333260536194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,6144,0.33027466138203937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,5120,0.00904444439543618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,5120,0.01759999990463257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,5120,0.3359991179572211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,5120,0.009648000200589498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,4096,0.007659555309348636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,7168,0.01075644459989336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,4096,0.016224000189039443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,3584,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,4096,0.33060089747111004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,4096,0.008655110994974772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,3584,0.01552088889810774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,3072,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,3584,0.008329778081840938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,3584,0.31714844703674316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,3072,0.014767110347747803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,2560,0.005964444329341252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,3072,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,3072,0.3329048951466878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,2560,0.014146667387750415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,2048,0.0053013331360287136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,2560,0.007631999750932057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,2560,0.31289421187506783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,2048,0.01351288871632682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,1536,0.0046640001237392426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,2048,0.007149332927332983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,2048,0.32656534512837726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,1536,0.013052444491121503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,1024,0.004268444246715969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,1536,0.006584888945023219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,1024,0.012720888687504662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,1024,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,768,0.0038951109680864546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,1536,0.4530835681491428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,1024,0.3231075604756673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,512,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,768,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,768,0.006301333506902059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,512,0.012224888636006249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,768,0.3372808827294244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,256,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,512,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,512,0.32427199681599933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,256,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,128,0.0028808888875775864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,256,0.0058666666348775225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,256,0.3255653381347656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,128,0.011798222031858234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,64,0.0028968888024489083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,1024,32,0.0031066667288541794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,1024,128,0.005633777628342311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,1024,128,0.29064710934956867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,65536,0.02946044339074029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,64,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,1024,32,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,16384,0.01595288846227858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,65536,0.1258808904223972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,65536,0.06002489063474867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,16384,0.03727022144529555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,65536,0.3643466631571452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,12288,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,16384,0.3224773406982422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,16384,0.017457778255144756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,12288,0.030789332257376775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,12288,0.014448000325096978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,10240,0.012525333298577202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,12288,0.31960444980197483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,10240,0.025789333714379206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,8192,0.010767999622556897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,10240,0.012813333008024426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,10240,0.35259021653069395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,8192,0.022707555029127333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,7168,0.009944889280531142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,8192,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,8192,0.3223297860887316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,7168,0.021294222937689886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,7168,0.01071999967098236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,7168,0.3563048839569092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,6144,0.018907555275493197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,6144,0.010096000300513374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,6144,0.3452586597866482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,5120,0.008833777573373582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,5120,0.01762488815519545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,5120,0.3338453239864773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,5120,0.009462222456932068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,4096,0.007908444437715743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,6144,0.009355555805895064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,4096,0.016208888755904306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,3584,0.0069715554515520734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,4096,0.3147679964701335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,3584,0.015508444772826301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,3584,0.3257315423753527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,3584,0.008300444318188561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,3072,0.006315555423498154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,3072,0.01479022204875946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,3072,0.3246408833397759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,3072,0.007948444121413762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,2560,0.01407199932469262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,4096,0.008678221868144141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,2560,0.007334222396214803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,2048,0.0053075556125905775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,2560,0.3297937711079915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,2048,0.013566222455766467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,2048,0.3254106574588352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,2048,0.007357333269384172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,1536,0.004918222212129169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,2560,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,1536,0.01311377767059538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,1024,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,1536,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,1536,0.33899821175469297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,768,0.0038720000949170855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,1024,0.006324444380071428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,1024,0.012802666260136498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,1024,0.33450934622022843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,768,0.01240888900227017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,512,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,768,0.006247111078765657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,768,0.3040657838185628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,512,0.012075555821259817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,256,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,512,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,512,0.3251840008629693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,256,0.011735999749766456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,256,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,128,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,256,0.3039786550733778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,64,0.002847111059559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,768,128,0.005807999935415056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,768,32,0.0028408887899584244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,768,128,0.2866391075981988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,64,0.011439111497667102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,65536,0.023727110690540735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,32,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,65536,0.059734225273132324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,65536,0.1255564424726698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,16384,0.012857777376969656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,65536,0.31870844629075795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,16384,0.03642933236228095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,16384,0.017172444197866652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,16384,0.2958470980326335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,12288,0.028839998775058325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,768,128,0.011397333608733283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,10240,0.00996977753109402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,12288,0.2954568862915039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,12288,0.014503111441930136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,10240,0.025750221477614507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,10240,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,10240,0.2947351137797038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,8192,0.02236266599761115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,8192,0.30134399731953937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,12288,0.011023110813564725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,7168,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,7168,0.02112355497148302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,7168,0.30370844735039604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,6144,0.0076142218377855085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,7168,0.010802666346232096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,6144,0.019122666782803006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,8192,0.00905600024594201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,6144,0.33261066012912327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,5120,0.00832266691658232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,6144,0.010031110710567897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,8192,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,5120,0.33143822352091473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,5120,0.009708444277445475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,4096,0.016157334049542744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,4096,0.35189421971638996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,4096,0.008650666309727563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,5120,0.01751733322938283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,4096,0.007655111451943715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,3584,0.015484443969196744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,3584,0.008078222473462423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,3072,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,3584,0.3351857662200928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,3072,0.014510222607188754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,3072,0.32979289690653485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,3072,0.007696000238259633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,2560,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,3584,0.00700800038046307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,2560,0.014179555906189812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,2048,0.0052284445199701525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,2560,0.007601777712504069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,2560,0.32675109969245064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,2048,0.013394667042626275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,1536,0.004561777744028303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,2048,0.007260444263617198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,2048,0.3420906596713596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,1536,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,1024,0.0038853333228164246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,1536,0.00693866651919153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,1536,0.32545600997077095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,1024,0.01239999963177575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,768,0.0038693332009845306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,1024,0.006599999964237213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,1024,0.33374667167663574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,768,0.012072888513406118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,512,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,768,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,512,0.011815110842386881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,512,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,512,0.3342426617940267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,256,0.0032222221295038858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,256,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,256,0.32410844167073566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,128,0.00314666661951277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,256,0.00601333338353369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,128,0.01144177797767851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,768,0.32943911022610134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,64,0.0028373334142896864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,512,128,0.005613333235184352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,512,32,0.0028400001012616684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,65536,0.019276445110638935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,64,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,512,32,0.011382222175598145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,65536,0.0529333319928911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,65536,0.325234678056505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,16384,0.009698666632175446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,16384,0.034972445832358465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,16384,0.3065982129838732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,12288,0.008299555215570662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,16384,0.01717511150572035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,512,128,0.31389954355027944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,12288,0.02881422307756212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,10240,0.009039999710188972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,12288,0.32862310939364964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,10240,0.025409777959187824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,65536,0.12527467144860163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,10240,0.012808000048001608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,8192,0.008325333396593729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,10240,0.319830232196384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,8192,0.022389334109094407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,7168,0.008300444318188561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,8192,0.3082871172163221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,7168,0.021018667353524104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,12288,0.014144000079896716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,7168,0.01069777790043089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,6144,0.007610666255156199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,6144,0.018907555275493197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,6144,0.3133431010776096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,6144,0.010009778042634329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,5120,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,8192,0.011392888923486074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,5120,0.3183928860558404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,5120,0.009568000005351173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,4096,0.007311999797821045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,4096,0.016187555260128446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,7168,0.310079124238756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,3584,0.007000000112586551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,4096,0.3200826644897461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,4096,0.008548444343937768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,3584,0.015537778536478678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,3072,0.006718222051858902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,3584,0.00828888929552502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,3584,0.3120826615227593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,3072,0.014789332946141561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,2560,0.005580444302823808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,5120,0.01754577789041731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,3072,0.3234142197502984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,2560,0.014174222946166992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,2560,0.305881765153673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,2560,0.007647111184067196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,2048,0.013538666897349887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,2048,0.35036622153388125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,3072,0.007999111380841997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,1536,0.00453244439429707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,2048,0.007243555453088548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,1536,0.013074666261672974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,2048,0.005223111146026188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,1024,0.003915555361244413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,1536,0.3333288828531901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,1536,0.006649777707126405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,1024,0.012402666939629449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,1024,0.0063075555695427795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,768,0.003548444559176763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,1024,0.33373154534233945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,768,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,512,0.003519111209445529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,768,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,768,0.3353751235538059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,512,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,256,0.003141333245568805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,512,0.005756444401211209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,512,0.3346577750311957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,256,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,128,0.0029075555503368378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,256,0.00563466673096021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,256,0.3325822353363037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,128,0.01146488885084788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,64,0.0028408887899584244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,256,128,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,256,32,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,256,128,0.3052542209625244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,65536,0.014548444085650973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,32,0.011425777441925473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,65536,0.12148355113135444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,65536,0.3003902170393202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,16384,0.009339555270142024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,65536,0.051402668158213295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,16384,0.03467288944456313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,16384,0.3327511151631673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,16384,0.016894222961531747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,12288,0.00832177781396442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,256,64,0.011678222152921887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,12288,0.028334223561816748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,10240,0.008655110994974772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,12288,0.014316444595654806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,10240,0.025331555141343012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,10240,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,10240,0.3227733241187202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,8192,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,8192,0.02236355510022905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,8192,0.3154462178548177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,7168,0.007662222617202335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,7168,0.020956445071432326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,12288,0.3287137879265679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,6144,0.007617777420414819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,7168,0.010817777779367236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,7168,0.3210444450378418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,6144,0.010035555395815108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,5120,0.006963555183675554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,6144,0.4470657772488064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,8192,0.01148977792925305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,5120,0.009576888548003303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,4096,0.00674222202764617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,5120,0.41573601298862034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,4096,0.016150222884284127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,6144,0.01888444523016612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,3584,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,4096,0.008691555923885768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,4096,0.3776177830166287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,3584,0.015574221809705099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,3072,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,3584,0.008345777789751688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,3584,0.3388213316599528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,5120,0.01719466679626041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,3072,0.01480888823668162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,2560,0.0063031112982167145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,3072,0.007876444194051955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,2560,0.014091556270917257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,3072,0.31748000780741376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,2048,0.006351999938488007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,2560,0.007599111232492659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,2560,0.3376053439246283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,1536,0.004586666822433472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,2048,0.013479111095269522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,2048,0.007152000235186682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,2048,0.3158782323201497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,1536,0.012809777425395118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,1024,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,1536,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,1536,0.31498222880893284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,1024,0.012290666500727335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,768,0.0038497779104444715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,1024,0.0062871111763848206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,1024,0.3137608899010552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,768,0.012084444363911947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,768,0.006275555739800136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,768,0.321831120385064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,512,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,512,0.34337422582838273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,256,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,256,0.01178311142656538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,256,0.3449689017401801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,512,0.0033395555284288195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,128,0.002833777831660377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,256,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,128,0.01144977741771274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,512,0.005920889063013925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,64,0.0025635556214385563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,192,128,128,0.005582222094138463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,192,128,128,0.32527732849121094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,128,32,0.0028239999794297745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,65536,0.014116444521480136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,64,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,16384,0.008725333544943068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,128,32,0.011443555355072021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,12288,0.008049777812427944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,65536,0.12053778436448838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,10240,0.007266666326257918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,16384,0.034324444002575345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,12288,0.028453333510292902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,8192,0.006940444310506185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,7168,0.006639110959238476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,10240,0.025378665990299646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,6144,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,7168,0.020779555042584736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,5120,0.006330666856633292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,6144,0.019146665930747986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,4096,0.005936000082227919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,5120,0.01723022262255351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,3584,0.006272888845867581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,4096,0.016141333513789706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,3072,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,3584,0.01512888901763492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,2560,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,3072,0.01478488908873664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,2048,0.006251555350091722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,2560,0.013759111364682516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,1536,0.004544888933499654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,2048,0.01350844403107961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,8192,0.022367111510700647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,1024,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,768,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,1536,0.012757333616415659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,512,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,1024,0.01260088880856832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,768,0.012400888734393649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,512,0.011728000309732227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,128,0.002893333426780171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,256,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,64,0.0027155555370781156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,128,0.011065777805116443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,32,0.0028906667398081887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,64,0.01166133334239324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,65536,0.01348622226052814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,64,32,0.010791999598344168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,16384,0.00833511104186376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,65536,0.11979466014438206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,12288,0.008004444340864817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,16384,0.034338666333092585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,10240,0.007274666594134436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,64,256,0.0029075555503368378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,12288,0.02881422307756212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,8192,0.006973333656787872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,7168,0.006589333216349284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,10240,0.025127111209763422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,6144,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,8192,0.022491556074884202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,5120,0.00628266649113761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,7168,0.02093688812520769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,4096,0.005954666684071223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,3584,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,5120,0.017536888519922893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,3072,0.006292444550328785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,4096,0.016203555795881484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,2560,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,3584,0.015264888604482016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,2048,0.006020444548792309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,2560,0.014038221703635322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,1536,0.004360888981156879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,2048,0.013596444494194455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,1536,0.013120888835854001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,1024,0.004283555679851108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,768,0.003918222255176968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,1024,0.012438221938080259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,512,0.003212444484233856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,768,0.012449777788586087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,256,0.0032151111712058387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,512,0.01201333353916804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,128,0.0029288888391521242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,256,0.011744889120260874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,64,0.0028320000403457214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,3072,0.014872888724009195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,128,0.011457777685589261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,192,32,32,0.0028408887899584244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,64,0.011501332951916588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,32,0.011411555939250521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,16384,0.34064356486002606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,16384,0.28593066003587514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,192,32,6144,0.018668444620238412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,12288,0.2591262128618029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,16384,0.13037867016262478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,12288,0.21749067306518555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,16384,0.3615146742926703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,10240,0.23619911405775282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,12288,0.3613777690463596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,10240,0.18523111608293322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,12288,0.0941457748413086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,10240,0.08134311437606812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,8192,0.1891608900494046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,8192,0.1548186673058404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,10240,0.3429528872172038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,8192,0.07070755296283297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,7168,0.16675199402703178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,7168,0.1293191115061442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,8192,0.33147022459242076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,6144,0.14417866865793863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,6144,0.11190222369299994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,7168,0.331986665725708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,6144,0.05199911197026571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,5120,0.12150400214725071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,7168,0.05870755513509115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,6144,0.32299378183152944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,5120,0.09591822491751777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,5120,0.04439289040035672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,4096,0.09954044553968643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,4096,0.08452177710003322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,5120,0.2921920087602404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,3584,0.08655911021762425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,3584,0.07219200001822577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,4096,0.329004446665446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,3072,0.07413599888483684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,3584,0.03341866532961527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,3584,0.31012267536587185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,3072,0.06299022171232435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,2560,0.06328088707394071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,3072,0.03016888764169481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,2560,0.05518044365776909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,4096,0.0374311109383901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,3072,0.3324524561564128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,2048,0.057370669311947294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,2560,0.026675555441114638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,2048,0.04855022165510389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,2560,0.31801867485046387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,1536,0.04025422202216254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,2048,0.02290488945113288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,1536,0.04101955559518602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,2048,0.3068106704288059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,1024,0.028499556912316218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,1536,0.0177831103404363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,1536,0.3341244326697455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,768,0.022646221849653456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,1024,0.014511111709806653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,1024,0.32996177673339844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,768,0.02882666720284356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,512,0.016016888949606154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,768,0.013400889105266996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,768,0.32762400309244794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,512,0.025054223007626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,256,0.01013777818944719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,512,0.012119111087587146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,512,0.3334026601579454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,256,0.022011554903454248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,128,0.008729777402347988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,256,0.011040888726711273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,256,0.32548801104227704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,1024,0.033082667324278094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,128,0.020356444848908316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,64,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,65536,32,0.008613333106040955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,65536,128,0.011032000184059143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,65536,128,0.30595021777682835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,64,0.020689777202076383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,65536,32,0.022103110949198406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,65536,0.3614213201734755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,16384,0.09715111388100518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,65536,0.35693955421447754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,65536,0.1433617803785536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,16384,0.09954133298661973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,12288,0.07510488563113742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,16384,0.04299022091759575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,16384,0.33026933670043945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,12288,0.07698844538794623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,10240,0.06298755274878608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,12288,0.03264888789918687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,12288,0.3300222290886773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,10240,0.0660239987903171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,8192,0.05119466781616211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,65536,0.43646399180094403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,10240,0.32585422197977704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,8192,0.057586669921875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,7168,0.0458551115459866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,8192,0.02513511147763994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,8192,0.33453954590691465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,7168,0.047804445028305054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,6144,0.039719998836517334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,7168,0.32790666156344944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,6144,0.042529778348075024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,10240,0.028696000576019287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,6144,0.3307982285817464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,5120,0.034690668185551964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,6144,0.018794667389657762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,5120,0.03839377893341912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,5120,0.33849867184956867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,4096,0.028057777219348486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,5120,0.01685244507259793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,7168,0.021171554923057556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,4096,0.034083555142084755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,4096,0.014630221658282809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,3584,0.025563556287023757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,4096,0.3441164493560791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,3584,0.03050666716363695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,3072,0.022319111559126113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,3584,0.3408062193128798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,3072,0.02681066592534383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,3072,0.32032354672749835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,3072,0.012480888929631976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,2560,0.024307555622524683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,2560,0.3348737822638617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,3584,0.01351111133893331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,2048,0.01717688971095615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,2560,0.01166133334239324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,2048,0.021907554732428655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,2560,0.019317333896954853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,2048,0.010370666782061258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,2048,0.3314097722371419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,1536,0.01960000064637926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,1536,0.32882844077216256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,1536,0.008612444831265343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,1024,0.010009778042634329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,1024,0.01757066614098019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,1024,0.3389155599806044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,768,0.008469333251317343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,768,0.01647111111217075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,1536,0.012830221818553077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,768,0.007316444483068254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,512,0.006623111251327727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,768,0.3266986740960015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,512,0.015171556009186639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,512,0.29425154791937935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,512,0.006924444602595435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,256,0.004575110971927643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,1024,0.007717333734035492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,256,0.014149333039919535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,128,0.004249777644872665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,256,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,256,0.31648532549540204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,128,0.013811555173661975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,64,0.004204444587230682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,16384,32,0.0041857775714662345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,16384,128,0.28915733761257595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,64,0.013925333817799887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,16384,32,0.014478221535682678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,65536,0.2631564405229357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,65536,0.2721653249528673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,16384,0.06936533583535089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,65536,0.1349804401397705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,16384,128,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,16384,0.07820444636874728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,12288,0.05509155657556322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,16384,0.04002133342954848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,16384,0.32300978236728245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,12288,0.06101510922114054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,10240,0.0469377769364251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,12288,0.030373334884643555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,12288,0.3199671109517415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,10240,0.053110222021738686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,10240,0.02626488937271966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,8192,0.039577778842714094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,10240,0.3281022177802192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,8192,0.046987556748920016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,7168,0.03460444344414605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,8192,0.02279111080699497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,8192,0.3436400095621745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,7168,0.039184000757005476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,6144,0.029976887835396662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,7168,0.019486222002241347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,7168,0.3258497714996338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,6144,0.03510311245918274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,5120,0.02607111136118571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,6144,0.017907554904619854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,6144,0.338027556737264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,5120,0.03130933311250474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,4096,0.021869333253966436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,5120,0.015992888145976596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,5120,0.32771023114522296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,4096,0.02716355522473653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,3584,0.019381332728597853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,4096,0.014089778065681458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,4096,0.3245626555548774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,3584,0.02437955637772878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,3072,0.016930666234758165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,3584,0.3231031099955241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,3072,0.022738666998015508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,3072,0.012146666646003723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,3072,0.3268817795647515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,2560,0.014861333701345654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,2560,0.021066667305098638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,2560,0.34540534019470215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,2560,0.01108088923825158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,65536,0.430913766225179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,2048,0.013912000589900546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,3584,0.013106666505336761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,2048,0.019458666443824768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,2048,0.009723555710580613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,1536,0.010777778095669217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,2048,0.33877865473429364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,1536,0.017535110314687092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,1536,0.008276444342401292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,1024,0.008431110945012834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,1536,0.3373146586947971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,1024,0.015631111131774057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,768,0.0074035558435651995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,1024,0.007626666790909237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,1024,0.3478231165144179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,768,0.014897776974572076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,512,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,768,0.007144889069928064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,768,0.3371733294592963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,512,0.014175110393100314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,256,0.004571555389298333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,512,0.0069724445541699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,512,0.3427075544993083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,256,0.013482666677898832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,128,0.003906666818592283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,256,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,256,0.3360124429066976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,128,0.013181333740552267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,64,0.0041759999261962045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,12288,128,0.006607999818192587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,12288,32,0.0038951109680864546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,12288,128,0.30751111772325307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,64,0.013355555633703867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,12288,32,0.013767110804716745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,65536,0.23676533169216582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,65536,0.2612657811906603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,16384,0.0647617777188619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,65536,0.1302782164679633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,16384,0.07630578014585707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,16384,0.03907910982767741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,16384,0.329403559366862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,12288,0.05045510994063484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,65536,0.4276017877790663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,12288,0.060600890053643115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,10240,0.043150222963756986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,12288,0.029046220911873713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,12288,0.3527928988138835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,10240,0.0518142216735416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,8192,0.03574488891495599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,10240,0.02570666703912947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,10240,0.31205066045125324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,8192,0.04530844423505995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,7168,0.03164444367090861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,8192,0.02182933357026842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,8192,0.33153067694769967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,7168,0.0373306671778361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,6144,0.027224888404210407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,7168,0.019201777047581144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,7168,0.3284906546274821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,6144,0.03362577822473314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,5120,0.023931556277804907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,6144,0.017608889275126986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,6144,0.3154097663031684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,4096,0.020650666620996263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,5120,0.029581334855821397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,5120,0.015825778245925903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,5120,0.2921537823147244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,4096,0.02643111182583703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,4096,0.013613333304723104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,4096,0.34067998992072207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,3584,0.023933332827356126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,3072,0.015511999527613321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,3584,0.3053724500868055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,3584,0.01275288893116845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,3072,0.022319111559126113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,2560,0.013813333378897773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,3072,0.011668444507651858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,3072,0.3082915676964654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,2560,0.020611556039916146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,2048,0.012520888613329994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,2560,0.01015377789735794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,2560,0.325034671359592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,2048,0.018985778093338013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,3584,0.01759466694460975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,1536,0.009947555760542551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,2048,0.00869599978129069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,2048,0.3201999929216173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,1536,0.01752000053723653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,1024,0.008010666403505538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,1536,0.007905777957704332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,1536,0.33511998918321395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,768,0.00655644428398874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,1024,0.0073315559162033936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,1024,0.32588622305128306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,768,0.014490667316648694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,512,0.005616000129116907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,768,0.007309333317809635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,768,0.33076000213623047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,512,0.014087999860445658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,512,0.33104533619350857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,256,0.004571555389298333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,256,0.013468444347381592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,1024,0.01514933341079288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,256,0.006632888896597757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,256,0.33562665515475804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,128,0.004257777912749184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,128,0.013080889152155982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,10240,128,0.3104800118340386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,128,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,64,0.012811555630630918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,32,0.0038773334688610504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,10240,32,0.013544888959990608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,65536,0.17218755351172554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,10240,64,0.003912000192536248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,65536,0.21921243932512072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,65536,0.08035022020339966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,16384,0.047981331745783486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,16384,0.06615644693374634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,65536,0.3437475628323025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,10240,512,0.006743999818960826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,12288,0.03850222296184964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,16384,0.025848888688617285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,16384,0.3424817721048991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,12288,0.05207911133766174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,12288,0.019243554936514962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,10240,0.03281955586539374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,12288,0.33151200082567006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,10240,0.04493688874774509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,8192,0.026259554757012263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,10240,0.01736177835199568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,10240,0.3383306662241618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,8192,0.03804444604449802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,7168,0.024145777026812237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,8192,0.015754666593339708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,8192,0.3004044426812066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,7168,0.03235822253757053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,6144,0.02179199953873952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,7168,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,7168,0.3270622094472249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,6144,0.02906044324239095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,5120,0.018225777480337355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,6144,0.012785777449607849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,6144,0.31564177407158744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,5120,0.025634666283925373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,4096,0.016415110892719693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,5120,0.3194168938530816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,4096,0.02360444433159298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,4096,0.010447999669445885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,4096,0.3203546735975477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,3584,0.013577777478430005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,3584,0.021698666943444148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,3072,0.012367111113336353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,3584,0.009159999589125315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,3584,0.31198665830824107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,3072,0.02028888960679372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,2560,0.011022222538789114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,3072,0.008644444247086843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,2560,0.018965333700180054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,3072,0.3986399968465169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,2048,0.01073333372672399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,2560,0.2986142105526394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,2560,0.007901333272457123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,2048,0.017300445172521803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,2048,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,1536,0.007994666695594788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,2048,0.300672001308865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,1536,0.01590844492117564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,1536,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,1024,0.0063493334584765965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,1536,0.36537777052985293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,1024,0.014496889379289416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,768,0.005247999810510212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,1024,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,1024,0.32758667733934194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,5120,0.01181422256761127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,768,0.014200000299347771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,512,0.004200889004601372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,768,0.00629333323902554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,512,0.01368711143732071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,768,0.3406177891625299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,256,0.0038444445365005066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,512,0.005954666684071223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,512,0.30251555972629124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,128,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,256,0.005993777679072485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,256,0.32658754454718697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,128,0.012721777790122561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,64,0.0035128887328836652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,8192,128,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,8192,128,0.3053342236412896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,8192,32,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,64,0.012730666332774691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,65536,0.15996000501844618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,65536,0.21694044272104898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,256,0.012792888614866467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,65536,0.07809066772460938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,16384,0.04435377650790744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,16384,0.06509955724080403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,65536,0.3118808799319797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,16384,0.3364444573720296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,8192,32,0.014121777481502958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,16384,0.024472888973024156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,12288,0.05052355594105191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,12288,0.3318328857421875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,12288,0.018569777409235638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,10240,0.030264887544843886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,10240,0.04384266667895847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,10240,0.33294932047526044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,10240,0.01662133302953508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,8192,0.025249777568711176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,12288,0.03547644284036424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,8192,0.03678399986690945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,7168,0.022487110561794702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,8192,0.32782578468322754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,8192,0.01481066644191742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,7168,0.03129422333505418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,6144,0.02036266691154904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,7168,0.013306666579511432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,7168,0.32793956332736546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,6144,0.02919911013709174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,5120,0.01722400055991279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,6144,0.012455110748608908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,6144,0.33915466732449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,5120,0.025770665870772466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,4096,0.014914666612943014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,5120,0.011384000380833944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,5120,0.3330266740587023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,4096,0.0232640008131663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,3584,0.012768000364303589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,4096,0.009853333234786987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,4096,0.3298799991607666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,3584,0.02130844526820713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,3072,0.011571555501884885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,3584,0.008970666262838576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,3584,0.33264001210530597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,3072,0.02021511064635383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,3072,0.008636444807052612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,2560,0.010406222608354356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,3072,0.3379848798116048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,2560,0.01891644464598762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,2048,0.010089777410030365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,2560,0.007640889121426477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,2560,0.3451804320017497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,2048,0.016890666551060148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,1536,0.007797333101431529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,2048,0.3305964469909668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,1536,0.015488889482286243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,1536,0.00700800038046307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,1536,0.32988354894849986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,1024,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,1024,0.014139556222491793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,1024,0.32075466050042045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,768,0.004878222114510006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,1024,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,2048,0.007316444483068254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,768,0.3365608851114909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,768,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,512,0.004934222333961063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,512,0.013441777891582914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,256,0.004280888785918553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,512,0.3240275647905138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,512,0.005994666781690385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,256,0.01274311128589842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,128,0.0038773334688610504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,256,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,256,0.3164489004347059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,128,0.012643555800120035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,64,0.0035093331502543557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,768,0.014097778333557976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,7168,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,7168,32,0.003682666768630346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,64,0.012774222426944308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,7168,32,0.01274222218328052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,65536,0.17896000544230142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,65536,0.07604444689220853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,16384,0.04826044374042087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,65536,0.33756266699896914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,7168,128,0.3044915464189318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,65536,0.147889773050944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,16384,0.022760000493791368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,12288,0.037392887804243304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,16384,0.32587377230326336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,12288,0.043755554490619235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,10240,0.03235022226969401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,12288,0.01789866718981001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,12288,0.33429598808288574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,10240,0.03634666734271579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,8192,0.026550221774313185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,10240,0.016188444362746347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,10240,0.3239697880215115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,8192,0.031332443157831825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,8192,0.014612444572978549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,7168,0.02331022255950504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,8192,0.32946756150987416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,7168,0.026994667119450037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,6144,0.02159022291501363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,7168,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,7168,0.34612356291876895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,6144,0.024290665984153748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,6144,0.30582666397094727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,5120,0.017619555195172627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,6144,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,5120,0.02274844381544325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,16384,0.05491822295718723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,4096,0.015296889675988091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,5120,0.3324088785383436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,5120,0.010626667075686984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,4096,0.020945777495702107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,3584,0.013516444298956128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,4096,0.31767021285163033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,3584,0.019482667247454327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,3584,0.008632000121805403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,3584,0.32134222984313965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,3072,0.011898666620254517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,3072,0.018198221921920776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,3072,0.3277840084499783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,3072,0.007976888782448238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,2560,0.010498666928874122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,4096,0.009373333719041612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,2560,0.016846223009957206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,2048,0.01035644445154402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,2560,0.007680888805124495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,2560,0.3168604373931885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,2048,0.01519733336236742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,2048,0.007296889192528195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,2048,0.32466310924953884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,1536,0.014929778046078153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,1536,0.00699733363257514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,1536,0.32663822174072266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,1024,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,1024,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,1024,0.30144532521565753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,768,0.005200888961553574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,1024,0.006346666564544042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,1536,0.007641777396202087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,768,0.3328088919321696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,768,0.006311111152172089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,512,0.004553777890072929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,512,0.012744888663291931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,512,0.32397598690456814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,256,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,256,0.012421333127551608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,768,0.013455111119482251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,256,0.33035466406080455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,128,0.0038542221817705366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,256,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,128,0.012164444559150271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,6144,128,0.3115831216176351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,128,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,6144,512,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,64,0.003573333223660787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,6144,32,0.0034959999223550162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,64,0.012177777787049612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,6144,32,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,65536,0.17685777611202666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,16384,0.04517333375083076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,65536,0.07213955455356173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,16384,0.05247199866506788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,65536,0.334861331515842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,16384,0.325055996576945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,12288,0.035487112071779035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,12288,0.040587554375330605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,12288,0.3076213200887044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,10240,0.03069066670205858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,12288,0.017439999514155917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,10240,0.034987555609809026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,10240,0.33688799540201825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,10240,0.015879111157523263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,8192,0.025174222058720056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,16384,0.021705778108702764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,8192,0.030856887499491375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,7168,0.02200977835390303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,8192,0.013836444252067141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,8192,0.3228462272220188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,7168,0.026757333013746474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,65536,0.13900089263916016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,7168,0.012755555411179861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,7168,0.31998221079508465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,6144,0.011884444289737277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,6144,0.3241724438137478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,5120,0.022164444128672283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,5120,0.3267937766181098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,6144,0.01975555552376641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,4096,0.014557333456145393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,5120,0.01035822182893753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,4096,0.020837333467271592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,6144,0.024322667055659827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,5120,0.01650399963061015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,3584,0.012688000169065265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,4096,0.3256595664554172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,4096,0.008813333180215623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,3072,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,3584,0.01921244462331136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,3584,0.008628444539176093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,3072,0.017942221628295053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,2560,0.010013333625263637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,3072,0.008027555214034187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,3584,0.40757065349155003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,3072,0.33455289734734434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,2560,0.016599999533759225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,2048,0.009200889203283522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,2560,0.008034666379292807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,2048,0.015191111299726697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,2048,0.30595821804470485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,1536,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,1536,0.01442311041884952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,1536,0.32727734247843426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,2560,0.3228026760949029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,1024,0.005992888990375731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,1536,0.0069155556460221606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,1024,0.013436444103717804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,2048,0.007281777759393056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,768,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,1024,0.33133776982625324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,1024,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,768,0.013426666458447775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,768,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,512,0.004589333302444882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,768,0.32440710067749023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,512,0.012789333032237159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,256,0.003913777983850903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,512,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,512,0.3220746782090929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,256,0.012129777835475074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,256,0.00563822231358952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,128,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,256,0.2919093237982856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,128,0.012423111332787408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,5120,128,0.005975111077229182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,5120,128,0.31150489383273655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,32,0.0031999999450312722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,64,0.01202044470442666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,5120,32,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,65536,0.09662133455276489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,65536,0.16712355613708496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,65536,0.06944533189137776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,65536,0.33637065357632107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,5120,64,0.003536888708670934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,16384,0.032623999648623996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,16384,0.05005422234535217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,16384,0.02094488839308421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,12288,0.028140445550282795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,16384,0.3261511060926649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,12288,0.03849777910444472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,10240,0.019983111156357657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,12288,0.017103999853134155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,12288,0.334715551800198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,10240,0.033246222469541765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,10240,0.01551466683546702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,8192,0.016522667474216886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,10240,0.3246817853715685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,7168,0.015082667271296183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,8192,0.029984000656339858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,7168,0.025798221429189045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,8192,0.4785466723971897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,7168,0.33105511135525173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,6144,0.013970666461520724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,6144,0.02397777802414364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,6144,0.011379555695586734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,5120,0.011542222566074796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,6144,0.4236275619930691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,5120,0.02200177808602651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,8192,0.01310400002532535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,4096,0.010339555641015371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,5120,0.009709333380063375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,5120,0.3848693370819092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,4096,0.019510222805870902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,7168,0.012097777591811286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,3584,0.009032888544930352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,4096,0.0087013335691558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,4096,0.32308443387349445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,3584,0.018273777431911893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,3072,0.008427555362383524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,3584,0.008344888687133789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,3584,0.3164569007025825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,3072,0.016840888394249808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,2560,0.007609777980380588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,3072,0.008062221937709385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,3072,0.3260408772362603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,2560,0.01588444411754608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,2560,0.007596444752481248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,2560,0.31500087844000924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,2048,0.015121777852376303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,1536,0.005211555709441503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,2048,0.007091555330488417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,2048,0.35393065876430935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,1536,0.014249778456158109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,1024,0.004270222038030624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,1536,0.006908444480763541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,1536,0.31612979041205513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,1024,0.013479111095269522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,768,0.004212444441186057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,1024,0.006263111200597551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,1024,0.3241964446173774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,768,0.013144888811641268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,512,0.0038888889054457345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,2048,0.007063999772071838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,768,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,768,0.3160604370964898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,512,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,256,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,512,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,512,0.29753245247734916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,256,0.0123831108212471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,256,0.00590133335855272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,128,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,256,0.3245991071065267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,128,0.012082666986518435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,64,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,4096,128,0.302635563744439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,4096,32,0.0031635556370019913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,64,0.01179377817445331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,4096,32,0.012424888710180918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,65536,0.08906933334138657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,65536,0.1667928828133477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,65536,0.32063200738694936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,65536,0.0690915584564209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,16384,0.029650668303171795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,4096,128,0.005660444498062134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,16384,0.04866755670971341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,12288,0.025585777229732935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,16384,0.020807999703619216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,12288,0.03790666659673055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,12288,0.31899110476175946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,12288,0.01683644453684489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,10240,0.024359999431504145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,10240,0.03336711062325372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,10240,0.3262408839331733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,10240,0.015371556083361307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,8192,0.015951999359660678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,16384,0.3218417697482639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,8192,0.3134746551513672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,7168,0.013817778064144982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,7168,0.02564711040920681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,7168,0.3043244414859348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,7168,0.01149866647190518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,6144,0.013732444081041547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,8192,0.029270221789677937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,6144,0.023582221733199224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,8192,0.013163555827405719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,5120,0.011785777906576792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,6144,0.3058577908409966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,6144,0.0107004443804423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,5120,0.021720000439220007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,4096,0.011746666497654386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,5120,0.009694221946928237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,5120,0.3020239935980903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,3584,0.010536889235178629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,4096,0.0199991116921107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,4096,0.008662222160233391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,4096,0.27422934108310276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,3072,0.008042666647169325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,3584,0.01832177738348643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,3584,0.00831822223133511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,3584,0.32086843914455837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,3072,0.016919111212094624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,3072,0.00794666674402025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,2560,0.007344000041484833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,3072,0.3197511037190755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,2048,0.006325333482689328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,2560,0.015853333804342482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,2560,0.007583111524581909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,2560,0.3192319869995117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,2048,0.015126221709781222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,1536,0.005233777893914117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,2048,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,2048,0.31961599985758465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,1024,0.004273777620659934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,1536,0.014488000008794995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,1536,0.0069191112286514705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,1536,0.3179866737789578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,1024,0.013468444347381592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,1024,0.32535378138224286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,768,0.013384000294738345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,768,0.006295111030340195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,512,0.003604444364706675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,768,0.2994951142205132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,512,0.012832000023788877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,512,0.30425333976745605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,256,0.003576000117593341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,256,0.01239822225438224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,768,0.003946666502290302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,256,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,128,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,256,0.3215137852562798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,128,0.011799111134476133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3584,128,0.295267555448744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,64,0.0028986665937635633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,512,0.00591822216908137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,64,0.011753777662913004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,1024,0.0063582224150498705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3584,32,0.01240533341964086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,65536,0.0725039972199334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,65536,0.12578577465481228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,65536,0.06803288724687365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,65536,0.32850755585564506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,16384,0.025388444463411968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,16384,0.038697779178619385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,16384,0.33065954844156903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,16384,0.02060622142420875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,12288,0.02216622233390808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,12288,0.030887110365761652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3584,128,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,12288,0.016568000117937725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,12288,0.3356871075100369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3584,32,0.0031991110493739447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,10240,0.021405332618289526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,10240,0.014807111687130399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,8192,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,10240,0.30853067504035103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,8192,0.02437866727511088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,8192,0.31665155622694224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,8192,0.012119111087587146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,7168,0.016047999262809753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,7168,0.021381333470344543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,7168,0.3212515513102214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,6144,0.014984000060293408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,10240,0.027095999982621934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,6144,0.0203155560625924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,6144,0.010392888552612728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,5120,0.012898666991127862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,6144,0.30526754591200084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,5120,0.018601778480741713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,5120,0.31164977285597056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,4096,0.010806221928861407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,5120,0.009677333136399588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,7168,0.010695999695195092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,4096,0.01700888905260298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,3584,0.009666666388511658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,4096,0.008640000389681922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,3584,0.01550044450494978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,3584,0.00829777783817715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,3584,0.3119057814280192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,3072,0.008837333156002892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,3072,0.014783999986118741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,3072,0.32939288351270885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,3072,0.007681777907742395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,2560,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,2560,0.014102222190962898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,4096,0.34177954991658527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,2560,0.31832265853881836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,2048,0.007213333414660559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,2560,0.007557333343558841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,1536,0.005658666706747479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,2048,0.013769778112570444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,2048,0.0070257774657673305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,2048,0.3108897738986545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,1536,0.013422222601042854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,1024,0.004878222114510006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,1536,0.31521066029866535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,1024,0.012856889102194043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,768,0.004566222015354368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,1024,0.3110355536142985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,768,0.012738666600651212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,768,0.006361777583758037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,512,0.0038622220357259116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,512,0.012483555409643384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,1536,0.006686222222116258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,512,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,512,0.34287556012471515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,256,0.003519111209445529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,1024,0.006351111249791251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,256,0.011907555162906647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,128,0.0032213332338465583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,256,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,256,0.31833865907457143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,768,0.32335111829969615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,128,0.011699555648697747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,64,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,3072,32,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,3072,128,0.30278754234313965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,64,0.011783999701340994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,65536,0.06386844317118327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,3072,32,0.011729777687125735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,65536,0.12302311261494954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,65536,0.0662248863114251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,16384,0.02307644486427307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,16384,0.036870221296946205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,65536,0.32675467597113717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,12288,0.020021332634819877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,16384,0.335664881600274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,16384,0.020290666156344943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,12288,0.030194666650560167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,10240,0.017423111531469557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,12288,0.016229333149062265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,12288,0.3215093347761366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,10240,0.027041777968406677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,8192,0.016544888416926067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,10240,0.014135999812020196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,10240,0.32589154773288304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,8192,0.02459555533197191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,8192,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,7168,0.014631999863518609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,8192,0.3322462240854899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,7168,0.021342222889264423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,6144,0.012734221915404001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,7168,0.011419555379284753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,7168,0.3247884379492866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,6144,0.019626667102177937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,3072,128,0.005571555760171678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,5120,0.011029333704047732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,6144,0.01036711119943195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,5120,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,5120,0.3086746798621284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,5120,0.009326222042242685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,4096,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,4096,0.016153777639071148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,4096,0.32097866800096303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,4096,0.008415111237102086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,3584,0.008997333546479544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,3584,0.015206222732861837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,6144,0.33348266283671063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,3072,0.008542222281297049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,3584,0.008316444853941599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,3072,0.014814222852389017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,3072,0.30170400937398273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,2560,0.007784888976150089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,3072,0.008021333151393467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,2560,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,2560,0.34152266714308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,2560,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,2048,0.013421333498424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,2048,0.3046533266703288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,2048,0.006981333096822103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,1536,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,1536,0.01311288856797748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,2048,0.00655733338660664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,3584,0.3397093349032932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,1536,0.006931555353932911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,1024,0.0052062223354975385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,1536,0.3312026659647624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,1024,0.012709333664841123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,1024,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,768,0.004934222333961063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,1024,0.3229600058661567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,768,0.012478222449620565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,512,0.0041964443193541635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,768,0.006059555543793573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,768,0.3395564556121826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,512,0.012147555748621622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,256,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,512,0.005865777946180767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,512,0.32344622082180446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,256,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,128,0.0035128887328836652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,256,0.0058782220714622075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,256,0.31371021270751953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,128,0.011438222395049201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,64,0.0031964443624019623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2560,128,0.005883555445406172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2560,32,0.00313688897424274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2560,128,0.2956622176700168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,65536,0.05278311173121134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,64,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2560,32,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,65536,0.12192532751295303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,16384,0.02104266650146908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,65536,0.06542666753133138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,65536,0.3293457825978597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,16384,0.03606399893760681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,12288,0.015824889143308003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,16384,0.019911110401153564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,12288,0.029230223761664495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,12288,0.016100444727473788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,10240,0.01444533301724328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,16384,0.45366668701171875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,12288,0.3212097750769721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,8192,0.013791110780504016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,10240,0.026380444566408794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,10240,0.013176889055305056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,10240,0.3266044457753499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,8192,0.02372177773051792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,7168,0.013364444176355997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,8192,0.011534222298198275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,8192,0.3153555658128527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,7168,0.02038755516211192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,6144,0.011899555722872416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,7168,0.01106844428512785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,7168,0.3180302249060737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,6144,0.01904355486234029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,6144,0.010085333552625444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,6144,0.3208391136593289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,5120,0.01701955497264862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,4096,0.009043555292818282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,5120,0.009667555491129557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,4096,0.015581332974963717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,4096,0.3269520070817736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,4096,0.008663111262851292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,3584,0.00795022232664956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,3584,0.014871999621391296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,3584,0.32904354731241864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,3584,0.008311111066076491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,3072,0.006934222247865465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,3072,0.014463111758232117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,3072,0.3352204428778754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,2560,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,5120,0.35233955913119847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,2560,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,2560,0.3510311179690891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,2560,0.007630222373538547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,5120,0.009992000129487779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,2048,0.00556355549229516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,2048,0.013464888764752282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,2048,0.006978666616810693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,2048,0.35042312410142684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,1536,0.004991111242108875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,1536,0.013208000196350945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,1536,0.3329431215922038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,1536,0.006716444260544247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,1024,0.01276622215906779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,1024,0.34944889280531144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,3072,0.007792000141408708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,768,0.0038559999730851916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,1024,0.00628977765639623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,768,0.012424888710180918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,1024,0.004543999830881755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,768,0.006251555350091722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,512,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,768,0.3415333429972331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,512,0.012079111403889127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,512,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,512,0.3218604458702935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,256,0.011771555576059552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,256,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,256,0.3261733320024279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,128,0.0031831111345026228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,128,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,2048,128,0.3149271011352539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,2048,128,0.005639111002286275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,64,0.0031991110493739447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,32,0.003173333282272021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,2048,256,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,64,0.011409777734014722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,65536,0.04141689009136624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,2048,32,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,65536,0.11949777603149414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,65536,0.06327466832266913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,16384,0.01573422220018175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,65536,0.32453155517578125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,16384,0.03567644291453891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,12288,0.014807111687130399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,16384,0.32600710127088756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,12288,0.029136889510684546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,12288,0.3278622097439236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,12288,0.014488889111412896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,10240,0.013840888937314352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,10240,0.02648711038960351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,8192,0.014389332797792224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,10240,0.3212062252892388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,8192,0.022330666581789654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,8192,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,16384,0.018958222534921434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,8192,0.32034754753112793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,7168,0.013189333180586496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,7168,0.01995022263791826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,7168,0.011014222270912595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,6144,0.010783111055692038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,7168,0.3042115635342068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,6144,0.018557333283954196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,6144,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,6144,0.3349440097808838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,5120,0.017148445049921673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,10240,0.013197333448463015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,4096,0.008351999852392409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,5120,0.009662222531106737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,5120,0.30582578976949054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,4096,0.01587466730011834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,4096,0.008707555631796518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,3584,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,4096,0.32198577457004124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,3584,0.014783999986118741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,3072,0.006837333242098491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,3584,0.008346666892369589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,3584,0.32713956303066677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,5120,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,3072,0.0143857780430052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,2560,0.006044444524579578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,3072,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,3072,0.3236595524681939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,2560,0.013807111316257052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,2048,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,2560,0.0075839997993575195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,2560,0.3316284550560845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,2048,0.013428444663683573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,1536,0.004967111266321606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,2048,0.007251555720965068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,2048,0.322946654425727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,1536,0.013173333472675748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,1024,0.004318222403526306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,1536,0.006696888970004187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,1536,0.32472088601854115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,1024,0.012392000191741519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,768,0.004016889052258598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,1024,0.006578666468461354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,1024,0.31932799021402997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,768,0.012123555772834353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,768,0.0063279999627007385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,768,0.3258151213328044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,512,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,512,0.2961120075649685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,512,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,256,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,256,0.011807111402352652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,256,0.3189146783616808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,256,0.005600888695981767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,128,0.0028924445311228433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,128,0.011437333292431302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,512,0.0036391110883818734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,64,0.0031519999934567344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1536,128,0.005578666511509154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1536,32,0.0034737777378824023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,64,0.011364444262451597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,65536,0.03076710965898302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1536,32,0.01105955574247572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,65536,0.11601421568128799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,65536,0.341830227110121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,65536,0.05983289082845052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,16384,0.035086221165127225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1536,128,0.3014266755845812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,16384,0.01848266687658098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,12288,0.014367999302016364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,16384,0.3027626673380534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,12288,0.028216001060273912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,12288,0.3157591025034587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,10240,0.012827555338541666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,12288,0.014505777094099255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,16384,0.016551999582184684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,10240,0.025633777181307476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,8192,0.011739555332395764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,10240,0.013071111506885953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,10240,0.34133866098192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,8192,0.022026666336589392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,7168,0.011130666567219628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,8192,0.011716444459226398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,8192,0.31401512357923717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,7168,0.019519999623298645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,7168,0.010718222293588849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,6144,0.009713778065310584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,7168,0.3451964325375027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,6144,0.01789777808719211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,5120,0.008999110923873054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,6144,0.010093332992659675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,6144,0.3360488944583469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,5120,0.01684977776474423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,4096,0.0075457774930530125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,5120,0.009683555199040307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,5120,0.3309333324432373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,4096,0.015455111861228943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,3584,0.010614222122563256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,4096,0.008671110702885522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,4096,0.32962221569485134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,3584,0.014820444915029736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,3072,0.009374221993817223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,3584,0.32330666648017037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,3072,0.014167111780908374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,3072,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,2560,0.008358221915033128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,2560,0.013762666947311826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,2560,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,2560,0.32783910963270396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,2048,0.007703999678293864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,3584,0.008336889247099558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,2048,0.013392888837390475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,2048,0.006966222491529252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,1536,0.00480266660451889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,1536,0.013068444199032254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,3072,0.32759467760721844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,2048,0.4467466672261556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,1024,0.004234666625658671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,1536,0.3200399875640869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,1536,0.006688888702127669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,768,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,1024,0.012852444416946836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,1024,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,1024,0.30665866533915204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,768,0.012446222205956778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,512,0.003576000117593341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,768,0.005988444305128521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,512,0.012114666402339935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,512,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,256,0.0033288889875014624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,512,0.3055555555555555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,256,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,128,0.003162666741344664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,256,0.005607111172543631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,256,0.3274915483262804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,128,0.01146133326821857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,64,0.0031999999450312722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,1024,128,0.005695111221737332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,1024,32,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,128,0.29929423332214355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,65536,0.025472000241279602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,64,0.011384888655609555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,1024,32,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,65536,0.11557155185275608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,16384,0.014111111561457316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,65536,0.29300355911254883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,1024,768,0.3891955481635199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,16384,0.0347324444188012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,16384,0.2975155512491862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,12288,0.013598221871587964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,16384,0.017214222086800467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,12288,0.028176890479193792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,12288,0.3408631218804254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,12288,0.014112888111008538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,10240,0.012075555821259817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,65536,0.059663110309176974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,10240,0.34748178058200413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,10240,0.013065777719020844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,8192,0.011140444212489657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,8192,0.02141422198878394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,8192,0.3462266657087538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,8192,0.011376000112957425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,7168,0.009709333380063375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,7168,0.019288000133302476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,7168,0.3364106549157037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,7168,0.010759111079904767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,10240,0.024031110935741003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,6144,0.009383999639087254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,6144,0.01824355622132619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,6144,0.01016711112525728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,5120,0.00872622181971868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,6144,0.3358782132466634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,5120,0.016908443636364408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,4096,0.007859555383523306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,5120,0.009332444104883406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,5120,0.3325715594821506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,4096,0.015792888071801927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,3584,0.0069875551594628235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,4096,0.008721777962313758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,4096,0.29714398913913304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,3584,0.014851555228233337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,3072,0.006462222172154321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,3584,0.33603289392259383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,3072,0.014501333236694336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,3072,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,3072,0.3346390989091661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,2560,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,2560,0.013834666874673633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,2560,0.34665777948167587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,2560,0.0074684446056683855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,2048,0.00554933316177792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,3584,0.008302222523424361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,2048,0.013427555561065674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,1536,0.00490133340160052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,2048,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,1536,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,1536,0.0069875551594628235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,1024,0.0038817777401871155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,1024,0.012477333347002665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,1024,0.3201564417945014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,768,0.0037360000941488477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,2048,0.32437777519226074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,1024,0.0064177778032090925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,768,0.012402666939629449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,512,0.003510222252872255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,768,0.0063031112982167145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,1536,0.3120008839501275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,512,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,256,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,512,0.00590133335855272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,256,0.01180622229973475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,512,0.4343048996395535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,128,0.003027555636233754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,256,0.30365512106153697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,256,0.005636444522274866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,768,128,0.005680888891220093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,128,0.011788444386588203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,64,0.003015111097031169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,768,32,0.0029057777590221832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,128,0.2800382243262397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,65536,0.01976355579164293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,64,0.011776000261306763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,768,32,0.011713777979214987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,768,768,0.30912711885240346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,16384,0.012609778179062737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,65536,0.11402844058142768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,65536,0.05903822183609009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,65536,0.3585057788425022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,16384,0.03427822060055203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,12288,0.01072177787621816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,16384,0.017214222086800467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,16384,0.2931973404354519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,10240,0.010024000373151567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,12288,0.026510222090615168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,12288,0.01426666643884447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,12288,0.33769954575432676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,10240,0.023716444770495098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,10240,0.01293155550956726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,10240,0.32696623272365993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,8192,0.02125866711139679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,7168,0.00832622249921163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,8192,0.011411555939250521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,8192,0.33043821652730304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,7168,0.019306666321224637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,6144,0.007686222592989604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,7168,0.01075466639465756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,7168,0.33598934279547793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,6144,0.01791733337773217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,6144,0.33226490020751953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,6144,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,5120,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,8192,0.008403555386596257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,5120,0.016851555969980028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,4096,0.007615110940403408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,5120,0.009655999640623728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,5120,0.33407023217942977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,4096,0.015494222442309061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,3584,0.00692622239391009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,4096,0.008647999829716152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,4096,0.3172871006859673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,3584,0.014797333214018079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,3072,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,3584,0.008299555215570662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,3584,0.3405786620246039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,3072,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,2560,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,3072,0.00794311116139094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,3072,0.3322933250003391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,2560,0.013751111096805997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,2048,0.0052595556610160405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,2560,0.0074728892909155945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,2560,0.33652623494466144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,2048,0.013096888860066732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,1536,0.004561777744028303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,2048,0.006993778049945831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,2048,0.3295564386579725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,1536,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,1024,0.004040889028045866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,1536,0.00693866651919153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,1536,0.33458222283257383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,1024,0.012368000215954252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,768,0.003847111016511917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,1024,0.006333333336644703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,1024,0.3036479949951172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,768,0.0121742222044203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,512,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,768,0.006217777729034424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,768,0.3275537755754259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,512,0.011759111450778114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,256,0.003212444484233856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,512,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,512,0.3155786726209852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,256,0.011715555356608497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,128,0.0028648889727062653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,256,0.005607111172543631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,256,0.3259066740671794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,128,0.011433777709801992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,64,0.0028684443483750024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,512,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,512,128,0.3025840123494466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,64,0.011212444139851464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,65536,0.0173333336909612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,512,32,0.011810666984981961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,65536,0.11246311664581299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,65536,0.3226524458991157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,16384,0.009678222239017487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,65536,0.05503377649519178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,16384,0.03289422061708238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,16384,0.3286071088578966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,512,32,0.002934222213096089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,12288,0.008842666943868002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,12288,0.026438222991095647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,12288,0.3085590998331706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,10240,0.01035644445154402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,12288,0.014176888598336114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,10240,0.023790222075250413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,10240,0.01310222182008955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,8192,0.011400889191362592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,10240,0.3037324481540256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,8192,0.02131555477778117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,7168,0.010430222584141625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,8192,0.011043555206722684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,16384,0.01724177764521705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,7168,0.01941066649225023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,6144,0.007730666961934831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,7168,0.01072088877360026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,7168,0.31394844584994847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,6144,0.018185777796639335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,5120,0.008332444561852349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,6144,0.010035555395815108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,6144,0.3261377811431885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,5120,0.016880000631014507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,4096,0.007643555601437886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,5120,0.00963644435008367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,5120,0.3160346613989936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,4096,0.015503111812803479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,3584,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,4096,0.00834222220712238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,8192,0.35816089312235516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,3584,0.01479288935661316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,3072,0.0069777775141927935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,3584,0.008274666965007782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,3584,0.3135840098063151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,3072,0.014257777068350049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,2560,0.005643555687533484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,3072,0.2893102169036865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,2560,0.013755555782053204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,2560,0.3117270999484592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,2560,0.007600000335110559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,4096,0.32138488027784556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,2048,0.004935111022657818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,2048,0.013487111363146039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,2048,0.3253137800428602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,3072,0.007627555893527136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,2048,0.007286222444640265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,1536,0.012808889150619507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,1536,0.33944445186191136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,1536,0.006664888726340399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,1024,0.003896000070704354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,1024,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,1024,0.335098663965861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,1024,0.006569777925809224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,768,0.003526222374704149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,768,0.012156444291273752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,1536,0.004582222137186262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,512,0.0032471112079090546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,768,0.31913601027594674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,768,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,512,0.012081777883900536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,512,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,256,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,512,0.31931376457214355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,256,0.01146044416560067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,128,0.0028391112056043413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,256,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,256,0.3272319899664985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,128,0.011380444798204633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,64,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,256,128,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,256,128,0.3009760114881727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,64,0.011083555718262991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,65536,0.014483556151390076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,256,32,0.011735999749766456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,65536,0.11021600166956584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,65536,0.3427120049794515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,16384,0.009398221969604492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,65536,0.05050844285223219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,16384,0.0331582228342692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,16384,0.33449334568447536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,256,32,0.002874666617976295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,16384,0.017256889078352187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,12288,0.008323555191357931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,12288,0.02646844420168135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,12288,0.014122666584120857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,10240,0.009033777647548253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,12288,0.32608622974819607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,10240,0.023452444208992854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,8192,0.008323555191357931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,10240,0.01310222182008955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,10240,0.3370533254411485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,8192,0.020961778031455144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,7168,0.007956444389290279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,8192,0.01148088855875863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,8192,0.32399733861287433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,7168,0.018894222047593858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,6144,0.007335999773608313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,7168,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,7168,0.3332737816704644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,6144,0.01755466726091173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,5120,0.006973333656787872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,6144,0.010107555323176915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,6144,0.327089786529541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,5120,0.016519111063745286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,4096,0.006665777828958299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,5120,0.009391110804345872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,5120,0.3241484430101183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,4096,0.015185778339703878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,3584,0.006616000086069107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,4096,0.008640000389681922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,4096,0.32335734367370605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,3584,0.014466666513019137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,3072,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,3584,0.00830577810605367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,3072,0.013845333622561561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,2560,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,3072,0.00794311116139094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,3584,0.4621306525336371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,2560,0.013439111411571503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,2048,0.006249777972698212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,2560,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,2560,0.34124088287353516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,2048,0.01311555587583118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,1536,0.004540444248252445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,2048,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,1536,0.012695999609099494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,2048,0.40900977452596027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,1024,0.003929777691761653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,1536,0.31723467508951825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,1536,0.006577777779764599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,1024,0.012447111308574677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,1024,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,768,0.0035786665976047516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,3072,0.3168231116400825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,1024,0.3155706723531087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,512,0.0034799998005231223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,768,0.011758222348160215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,768,0.006083555519580841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,512,0.012056888805495368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,512,0.32620710796780056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,512,0.005946666830115848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,256,0.0031928889867332247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,256,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,256,0.29557598961724174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,128,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,128,0.01109866632355584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,768,0.3564008871714274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,160,128,128,0.2995084391699897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,64,0.0028942223224374983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,128,0.00563466673096021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,128,32,0.0028391112056043413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,64,0.010779555473062726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,65536,0.01551999979548984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,128,32,0.011408000356621213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,16384,0.008984888593355814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,160,128,256,0.005672889037264719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,12288,0.008279110822412703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,65536,0.11003644598854913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,10240,0.007358222372002072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,16384,0.03219822380277846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,8192,0.007258666886223688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,12288,0.02647466626432207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,7168,0.0069884442620807225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,8192,0.021355556117163763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,6144,0.0063635553750726916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,7168,0.019256888164414298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,5120,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,6144,0.017989334132936265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,4096,0.005893333504597346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,5120,0.01685333251953125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,3584,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,4096,0.015441776977645027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,3072,0.006054222169849608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,3584,0.01481244464715322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,2560,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,3072,0.014149333039919535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,2048,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,2560,0.013793778088357715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,10240,0.023378666904237535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,1536,0.004500444564554426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,1024,0.003906666818592283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,2048,0.013131555583741931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,1536,0.01273333364062839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,768,0.0038213332494099936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,512,0.003198222153716617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,1024,0.012455999851226807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,256,0.0031600000543726813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,768,0.011741333537631564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,128,0.0028844444702068963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,512,0.012055111428101858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,64,0.0028737777223189673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,256,0.011783999701340994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,64,32,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,128,0.01109955542617374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,65536,0.013434666726324292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,64,0.011056889262464313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,16384,0.008681777450773451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,64,32,0.01146133326821857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,12288,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,65536,0.10884266429477268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,16384,0.032256000571780734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,10240,0.007334222396214803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,12288,0.026521777113278706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,8192,0.007261333366235097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,7168,0.006687110910813014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,10240,0.023715555667877197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,8192,0.02126666737927331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,7168,0.018915555543369718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,5120,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,6144,0.018197332819302876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,4096,0.005944889038801193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,5120,0.016344888342751395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,3584,0.006266666783226862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,4096,0.01555022266176012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,3072,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,3584,0.014487110906177096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,2560,0.006607999818192587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,3072,0.014173333843549093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,2048,0.006257777826653586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,2560,0.013453333742088742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,1536,0.004540444248252445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,1024,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,2048,0.013432000246312885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,6144,0.0063271112740039825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,1024,0.012139555480745105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,1536,0.013145777914259167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,512,0.003181333343187968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,256,0.003154666680428717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,768,0.012106666962305704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,128,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,512,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,64,0.0028248888750871024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,256,0.011365333365069495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,32,0.00254488881263468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,128,0.011685333318180509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,64,0.0110871113008923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,160,32,32,0.010417777631017896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,16384,0.358641783396403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,16384,0.21995111306508383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,160,32,768,0.003649777836269803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,12288,0.2688364452785916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,16384,0.1402746703889635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,12288,0.17162488566504586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,12288,0.10390755865308975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,16384,0.3470284408993191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,10240,0.2316382196214464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,10240,0.14217155509524876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,12288,0.3466133276621501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,10240,0.08751289049784343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,8192,0.1935137775209215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,8192,0.12018844816419814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,10240,0.34112710422939724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,8192,0.07895288864771526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,7168,0.16593510574764675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,8192,0.3362409008873834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,7168,0.10149777597851223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,6144,0.1452933284971449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,7168,0.06484799914889865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,6144,0.08810666534635757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,7168,0.35331111484103733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,6144,0.05445511142412821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,5120,0.12073066499498154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,5120,0.07699910799662273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,6144,0.3407626681857639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,4096,0.1007759968439738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,5120,0.04746133420202467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,4096,0.07034755415386625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,5120,0.3370373249053955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,3584,0.08667377630869548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,4096,0.04063911239306132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,4096,0.33123734262254506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,3584,0.05813688702053494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,3072,0.07437777519226074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,3584,0.038218667109807335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,3584,0.31636089748806423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,3072,0.03310488992267185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,3072,0.33129776848687065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,2560,0.06387288702858819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,2560,0.045801778634389244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,2048,0.051085333029429116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,2560,0.027832888894610937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,2560,0.31973245408799916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,2048,0.041073779265085854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,1536,0.04018400112787882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,2048,0.02360800074206458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,3072,0.05208711160553826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,1536,0.03466044531928168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,2048,0.41367289755079484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,1536,0.019309333629078336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,1536,0.2948373423682319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,1024,0.027909331851535376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,768,0.022250667214393616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,1024,0.014854222536087036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,1024,0.34128710958692765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,768,0.023235556152131822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,512,0.015263110399246216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,768,0.013473778135246701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,768,0.30748266643948025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,512,0.02015999952952067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,256,0.01036977767944336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,512,0.012138666378127204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,512,0.3334328863355849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,256,0.018575999471876357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,1024,0.028107557031843398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,256,0.01096088853147295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,256,0.3144551118214925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,128,0.017864889568752713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,64,0.007688889073001013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,65536,128,0.01108444482088089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,32,0.008633777499198914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,65536,128,0.2981751230027941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,64,0.019575110740131803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,65536,32,0.023137778043746948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,65536,0.29986222585042316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,65536,0.271326224009196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,16384,0.0800853305392795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,65536,128,0.00777599960565567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,16384,0.07964889208475749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,65536,0.15939999951256648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,12288,0.06247466802597046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,16384,0.046302222543292575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,12288,0.06408711274464925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,12288,0.03437066740459866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,10240,0.053378668096330434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,12288,0.33595556683010525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,65536,0.3308115535312229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,10240,0.05420710974269443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,8192,0.043528887960645885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,10240,0.032855110036002264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,10240,0.32557956377665204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,8192,0.045904888047112354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,7168,0.038650665018293597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,16384,0.30800533294677734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,8192,0.026872888207435608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,8192,0.3205511040157742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,6144,0.033815112378862165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,7168,0.022481777601771887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,7168,0.30482400788201225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,6144,0.03587288988961114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,6144,0.020976889464590285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,6144,0.38122399648030597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,5120,0.029275556405385334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,5120,0.03232977787653605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,5120,0.3375244405534532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,4096,0.024365334047211543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,5120,0.017881777551439073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,4096,0.02918222215440538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,4096,0.33610577053493923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,3584,0.022421333524915908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,4096,0.01607644392384423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,3584,0.02567911148071289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,3584,0.34519023365444607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,3584,0.015232000086042615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,3072,0.01925511161486308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,7168,0.04002755549218919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,3072,0.023384888966878254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,2560,0.01664355562792884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,3072,0.32396888732910156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,3072,0.012802666260136498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,2560,0.021371554997232225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,2560,0.012179555164443122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,2560,0.32336799303690594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,2048,0.019621334142155118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,2048,0.010870222416188983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,2048,0.32050755288865834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,1536,0.011328000161382886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,1536,0.01756266587310367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,1536,0.3386373254987929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,1536,0.009722666607962714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,1024,0.00851111114025116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,2048,0.015816888875431485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,1024,0.01622577839427524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,768,0.007620444728268518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,1024,0.3152302106221517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,1024,0.007962666451931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,768,0.014496000276671516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,512,0.005934222290913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,768,0.0075937774446275495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,768,0.3203626738654243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,512,0.013669333524174161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,256,0.004273777620659934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,512,0.006989333364698622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,512,0.3188657760620117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,256,0.013131555583741931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,128,0.003947555604908201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,256,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,256,0.31837866041395396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,128,0.012770666844315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,64,0.0038702223036024305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,16384,128,0.0069511110583941145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,16384,32,0.003925333420435588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,64,0.013426666458447775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,16384,32,0.014927110738224454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,65536,0.2277439965142144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,65536,0.21589510970645479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,16384,0.06202666627036201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,16384,128,0.30714578098720974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,65536,0.14089155197143555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,16384,0.0655075576570299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,12288,0.04841422372394138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,16384,0.04037244452370538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,12288,0.05384177631802029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,16384,0.31796354717678493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,65536,0.3295208877987332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,10240,0.041259557008743286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,12288,0.03155111273129781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,12288,0.3368142180972629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,10240,0.04607022139761183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,8192,0.034512887398401894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,10240,0.026929777529504564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,10240,0.317720890045166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,8192,0.03917600048912896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,7168,0.030440890126758154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,8192,0.023549333214759827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,8192,0.3312968942854139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,6144,0.027169777287377253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,7168,0.020170667105250888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,7168,0.31239112218221027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,6144,0.0307004451751709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,5120,0.023319111929999456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,6144,0.018263111511866253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,6144,0.30880533324347603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,5120,0.02752888864941067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,5120,0.016534222496880423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,4096,0.019655999210145738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,5120,0.3287271128760444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,4096,0.024546666277779475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,3584,0.01752355529202355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,4096,0.014451555079884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,4096,0.32273954815334743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,7168,0.034020443757375084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,3072,0.01516888870133294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,3584,0.021929777330822412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,3584,0.013256000147925483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,3584,0.32113154729207355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,3072,0.0203155560625924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,3072,0.012466666599114736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,3072,0.3294746610853407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,2560,0.01921955578856998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,2560,0.01147555559873581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,2560,0.3347146775987413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,2048,0.012902222573757172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,2048,0.018168000711335074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,2048,0.329968876308865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,2048,0.009844444692134857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,1536,0.009677333136399588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,2560,0.013592888911565145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,1536,0.3254657851325141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,1024,0.007369777394665613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,1024,0.015005333556069268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,1024,0.31732000244988334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,1024,0.007638221813572778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,768,0.0069751110341813825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,1536,0.016554666890038382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,768,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,1536,0.008664000365469191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,512,0.004945777770545748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,768,0.3071724573771159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,768,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,512,0.013182222015327878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,256,0.004571555389298333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,512,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,512,0.3135297828250461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,256,0.012728888955381183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,128,0.003903999924659729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,256,0.006629333313968446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,256,0.32731199264526367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,128,0.01273155543539259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,64,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,12288,128,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,12288,32,0.003863111138343811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,12288,128,0.31004354688856334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,64,0.01276622215906779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,12288,32,0.013169777890046438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,65536,0.19029688835144043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,65536,0.21515377362569174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,16384,0.052017778158187866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,65536,0.1333964400821262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,16384,0.06448977523379855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,16384,0.03995111253526475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,12288,0.04109066724777222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,16384,0.3418240017361111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,65536,0.3307866785261366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,12288,0.05320444371965197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,10240,0.03580355644226074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,12288,0.030722667773564655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,12288,0.328858666949802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,10240,0.04555022385385302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,8192,0.028941333293914795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,10240,0.02586311101913452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,10240,0.3265804449717204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,8192,0.03808710972468058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,7168,0.025926222403844196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,8192,0.023136888941129048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,8192,0.33204176690843373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,7168,0.03332711259524027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,6144,0.022812444302770827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,7168,0.019704000817404855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,6144,0.028982222080230713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,5120,0.019863999552196927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,6144,0.018248889181349013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,5120,0.026132444540659588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,6144,0.46468533409966367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,4096,0.017722666263580322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,5120,0.016050666570663452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,5120,0.3038311004638672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,4096,0.013799111048380533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,4096,0.024384000235133704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,3584,0.014930667148696052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,4096,0.32935023307800293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,3584,0.021536888347731695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,3072,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,3584,0.013791999883121915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,7168,0.30983734130859375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,3584,0.33741066190931535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,3072,0.020488889680968392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,2560,0.01182666669289271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,3072,0.01200977795653873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,3072,0.32768532964918345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,2048,0.011063111325105032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,2560,0.010679999987284342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,2560,0.31543289290534127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,2048,0.017875555488798354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,1536,0.008664888640244802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,2048,0.0094373333785269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,2048,0.3093706766764323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,1536,0.016019556257459853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,1536,0.008535111116038429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,1536,0.3365608851114909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,1024,0.014132445057233175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,1024,0.007639110916190677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,768,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,1024,0.322197331322564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,2560,0.018904000520706177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,768,0.013392888837390475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,512,0.004587555511130227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,768,0.007256000406212277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,768,0.32585512267218697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,512,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,512,0.29447910520765513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,1024,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,256,0.004212444441186057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,256,0.01277599980433782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,256,0.31350578202141655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,256,0.006581333362393909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,128,0.0038684445122877755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,128,0.012416889270146688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,10240,128,0.28739823235405815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,64,0.003555555724435382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,128,0.00664444433318244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,10240,32,0.0038577777643998465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,64,0.012479110724396177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,10240,512,0.00702044450574451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,10240,32,0.013432000246312885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,65536,0.17366311285230848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,65536,0.1737742159101698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,16384,0.04221955604023404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,65536,0.07889422443177965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,16384,0.052645332283443876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,16384,0.02509244448608822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,12288,0.03433777888615926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,16384,0.33406310611301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,12288,0.04347999890645345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,12288,0.3281795448727078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,12288,0.018942221999168396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,10240,0.02941511074701945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,10240,0.03766399953100417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,10240,0.016953777935769822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,8192,0.024243555135197107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,65536,0.3447813457912869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,8192,0.030961778428819444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,8192,0.015298666225539314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,8192,0.3332177797953288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,7168,0.021935110290845234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,7168,0.028047998746236164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,7168,0.32001956303914386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,6144,0.019553777244355943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,7168,0.013527111046844058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,10240,0.3258151213328044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,6144,0.024897777371936377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,5120,0.0165955556763543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,6144,0.3284168773227268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,5120,0.022442667020691767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,5120,0.011500444677140979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,5120,0.3277137809329563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,4096,0.014223999447292753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,4096,0.021624000536070928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,4096,0.3250613212585449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,4096,0.009908444351620143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,3584,0.019208000765906442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,6144,0.012423999607563019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,3584,0.3295057879553901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,3584,0.009425777528021071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,3072,0.018608889646000333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,3072,0.3188657760620117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,2560,0.009880888793203566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,3072,0.00870577742656072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,2560,0.017417778571446735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,3584,0.013170666992664337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,2048,0.00980355590581894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,2560,0.3339075512356228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,2560,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,2048,0.016657777958446078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,2048,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,1536,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,3072,0.011355555719799466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,2048,0.3153475655449761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,1024,0.0064791109826829694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,1536,0.014449778530332776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,1536,0.006928000185224745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,1024,0.013761777844693927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,1536,0.3267973264058431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,768,0.00490399988161193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,1024,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,1024,0.31725600030687123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,768,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,512,0.004270222038030624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,768,0.006242666807439592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,768,0.31826310687594944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,512,0.012763555679056378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,256,0.0038853333228164246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,512,0.005976888868543837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,512,0.3187982241312663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,128,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,256,0.005958222266700532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,256,0.3178657690684001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,128,0.011807111402352652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,64,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,8192,128,0.005934222290913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,8192,128,0.29514577653672963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,8192,32,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,64,0.01201244443655014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,32,0.012390221986505719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,65536,0.15347199969821507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,65536,0.16764267285664877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,16384,0.03827733463711209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,65536,0.07650666766696505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,8192,256,0.012317332956526013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,16384,0.051573332813051015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,16384,0.02383377816942003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,16384,0.3385804494222005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,65536,0.2897368801964654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,12288,0.0304115547074212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,12288,0.04251555601755778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,12288,0.018260444204012554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,10240,0.02622311148378584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,12288,0.303600894080268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,10240,0.03559911251068115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,8192,0.021272889441914026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,10240,0.01629244453377194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,10240,0.339397324456109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,8192,0.03028622269630432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,7168,0.01945688823858897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,8192,0.014850666125615438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,8192,0.3232373396555583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,7168,0.026325333449575637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,6144,0.017471111483044095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,7168,0.013191111385822296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,7168,0.3392328951093886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,6144,0.024354666471481323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,5120,0.014536889062987434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,6144,0.012339555554919772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,6144,0.3248426649305555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,5120,0.022325333621766832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,4096,0.012855999999576144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,5120,0.011364444262451597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,5120,0.32170844078063965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,4096,0.02100533412562476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,3584,0.011518222590287527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,4096,0.3269146548377143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,3584,0.018764444523387484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,3584,0.008754666480753157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,3584,0.3178595436943902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,3072,0.010120888551076254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,3072,0.018260444204012554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,3072,0.3312408924102783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,2560,0.009516444471147325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,3072,0.00832799987660514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,4096,0.009663999908500248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,2560,0.016920000314712524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,2048,0.008935111264387766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,2560,0.3368657694922553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,2560,0.007701333198282454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,2048,0.015470221638679504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,1536,0.0069164443347189165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,2048,0.007306666837798224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,2048,0.3271191120147705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,1536,0.013965333501497904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,1024,0.004912000149488449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,1536,0.0070257774657673305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,1536,0.3302728864881727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,1024,0.013437333206335703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,768,0.004562666846646203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,1024,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,1024,0.32680177688598633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,768,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,512,0.003921777837806278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,768,0.006232000059551663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,768,0.32396088706122506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,512,0.012734221915404001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,256,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,512,0.0059368887709246735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,512,0.32235556178622776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,256,0.012123555772834353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,128,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,256,0.32022754351298016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,128,0.011707555916574268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,256,0.006005333529578314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,7168,128,0.2936640050676134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,7168,128,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,32,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,64,0.012115555504957834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,65536,0.1384755505455865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,65536,0.16369688510894775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,65536,0.07462400197982788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,16384,0.03530044356981913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,7168,64,0.0032417778339650896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,65536,0.3361990981631809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,16384,0.05044888787799411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,12288,0.02769333289729224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,16384,0.022328888376553852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,7168,32,0.012773333324326409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,16384,0.29288087950812447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,12288,0.040748443868425153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,12288,0.017625777257813346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,10240,0.03452622228198581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,12288,0.3260444535149468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,8192,0.01995733380317688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,10240,0.015853333804342482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,10240,0.32936178313361275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,8192,0.029184887806574505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,7168,0.01773599949147966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,8192,0.014470222923490735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,8192,0.33243023024664986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,7168,0.025986666480700176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,6144,0.015842666228612263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,7168,0.012763555679056378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,7168,0.3264728917015923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,10240,0.023781332704755995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,6144,0.023765333824687537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,5120,0.013575110998418597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,6144,0.011923555698659686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,6144,0.32924267980787486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,5120,0.02217866645918952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,4096,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,5120,0.010741333166758219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,5120,0.31659555435180664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,4096,0.020928889513015747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,3584,0.010760889285140567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,4096,0.009323555562231276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,4096,0.33870400322808164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,3584,0.01903288894229465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,3072,0.009715555442704095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,3584,0.008708444734414419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,3584,0.3186577690972222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,3072,0.017864000466134813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,2560,0.008986666798591614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,3072,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,3072,0.3431377675798204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,2560,0.016523554921150208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,2048,0.0086986662613021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,2560,0.3022150993347168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,2560,0.007602666815121968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,2048,0.015465777781274585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,1536,0.006523555351628198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,2048,0.0069919998447100324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,2048,0.3092506726582845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,1536,0.014113777213626437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,1024,0.005101333475775189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,1536,0.0069475554757648045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,1536,0.3419991069369846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,1024,0.013445333474212222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,768,0.004287999951177173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,1024,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,1024,0.3188835514916314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,768,0.012905778156386482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,512,0.00416088890698221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,768,0.006258666515350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,768,0.3320462173885769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,512,0.012747555143303342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,256,0.003887999802827835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,512,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,512,0.281583997938368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,256,0.012069332930776807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,128,0.003666666646798452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,256,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,256,0.3223715623219808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,128,0.012087110843923358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,64,0.003587555554178026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,6144,128,0.005969777703285217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,6144,128,0.28950844870673287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,6144,32,0.0036142224238978494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,64,0.01201600001917945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,65536,0.11172978083292644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,65536,0.15959289338853624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,65536,0.0715199973848131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,16384,0.03679733475049337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,65536,0.3420844342973497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,16384,0.049439999792310924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,6144,32,0.011739555332395764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,16384,0.32938488324483234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,16384,0.02102311121092902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,12288,0.03127110997835795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,12288,0.0388026667965783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,10240,0.022388445006476507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,12288,0.3058302137586805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,10240,0.03334577878316244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,10240,0.3126346535152859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,10240,0.015466666883892484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,8192,0.018533334136009216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,8192,0.02846222122510274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,7168,0.016526222229003906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,8192,0.31004089779324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,8192,0.013814222481515674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,7168,0.02549955579969618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,12288,0.017218665944205392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,6144,0.015235554840829639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,7168,0.0125591109196345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,7168,0.3017768859863281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,5120,0.012772444221708508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,6144,0.023451555106374953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,6144,0.01145600030819575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,6144,0.3385653230879042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,5120,0.02199111051029629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,4096,0.01161688897344801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,5120,0.010129777921570672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,5120,0.33041244082980686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,3584,0.009526222116417354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,4096,0.008456000023418004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,4096,0.32441067695617676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,3584,0.018558222386572096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,3072,0.008746667040718926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,3584,0.008004444340864817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,3072,0.017157332764731515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,3584,0.3710791005028619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,2560,0.008216000265545314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,3072,0.33767112096150714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,3072,0.008006222546100616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,2560,0.016201777590645682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,2560,0.33454399638705784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,2560,0.007603555917739868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,2048,0.007666666474607255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,2048,0.014888000157144336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,2048,0.3320053418477376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,2048,0.007294221884674496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,1536,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,1536,0.014121777481502958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,1536,0.33163287904527455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,1536,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,4096,0.02058844433890449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,1024,0.013449778159459433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,1024,0.00628977765639623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,768,0.004585777719815572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,1024,0.33977601263258195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,768,0.0129466669427024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,768,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,512,0.0038933331767717996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,1024,0.004594666676388847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,512,0.012476444244384766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,512,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,512,0.3362702263726129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,256,0.0038577777643998465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,256,0.012135110795497894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,256,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,128,0.003540444291300244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,768,0.33647823333740234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,128,0.011778666741318174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,128,0.30025243759155273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,5120,128,0.005971555494599872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,64,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,5120,32,0.003516444315512975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,64,0.01240355521440506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,5120,32,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,65536,0.09394222497940063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,5120,256,0.34624356693691677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,65536,0.11549422476026748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,16384,0.03081688947147793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,65536,0.06894222233030531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,16384,0.03769866625467936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,65536,0.34887199931674534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,12288,0.025936888323889837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,16384,0.020473778247833252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,16384,0.3471697701348199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,12288,0.03087110983000861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,10240,0.019288889235920377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,12288,0.016712889075279236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,12288,0.32748177316453725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,10240,0.02643644478585985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,8192,0.015586665934986539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,10240,0.015068444940778943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,10240,0.3313431209988064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,8192,0.023335999912685815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,7168,0.014256889621416727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,8192,0.01312888910373052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,8192,0.33651823467678493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,7168,0.02104266650146908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,6144,0.013674666484196981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,7168,0.01183644433816274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,7168,0.3305484453837077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,6144,0.019811555743217468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,6144,0.011034666664070554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,6144,0.3204808764987522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,5120,0.01830311119556427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,4096,0.010384889112578498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,5120,0.31783821847703725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,4096,0.016695110334290397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,4096,0.3239040109846327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,4096,0.008316444853941599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,3584,0.008751110898123847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,3584,0.015504888362354703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,3584,0.3177937666575114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,3072,0.008045333127180735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,3584,0.008303110798199972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,3072,0.015240889456537036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,3072,0.3404248820410834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,3072,0.007651555869314406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,2560,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,2560,0.014477334088749357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,2560,0.3271137873331706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,2560,0.007337777978844113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,2048,0.006501333167155583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,5120,0.011367110742463006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,2048,0.013783111340469785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,1536,0.005457777943876054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,2048,0.3004319932725695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,2048,0.007000000112586551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,5120,0.010016000105275048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,1536,0.013108444710572561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,1024,0.004334222111437055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,1536,0.00664444433318244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,1536,0.3063111040327284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,1024,0.012749333348539142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,768,0.003910222401221593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,1024,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,1024,0.3067493438720703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,512,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,768,0.012464000119103326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,768,0.31253690189785427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,512,0.012392000191741519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,512,0.3376159932878282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,256,0.003306666596068276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,256,0.011798222031858234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,256,0.33993511729770237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,256,0.005904888941182031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,128,0.0031884445084465873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,128,0.011763555308183035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,768,0.006246222390068903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,64,0.00350577798154619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,4096,128,0.005664000080691443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,4096,32,0.0034186666210492453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,64,0.011839110818174152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,4096,32,0.011774222056070963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,65536,0.08462844292322795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,65536,0.11732533242967393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,65536,0.06795377863778009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,65536,0.3499644332461887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,16384,0.026479111777411565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,4096,128,0.3043564425574409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,16384,0.03623022304640876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,16384,0.020430222153663635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,12288,0.02230311102337307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,16384,0.331608878241645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,12288,0.029450668228997126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,10240,0.019302222463819716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,12288,0.34465421570671934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,10240,0.026032888227038916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,10240,0.3388053311241998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,10240,0.014874666929244995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,8192,0.015509333875444202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,8192,0.02274311085542043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,8192,0.3321617709265815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,8192,0.012460444536474017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,7168,0.013296888934241401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,12288,0.01664533383316464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,7168,0.3422559897104899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,7168,0.01147466649611791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,6144,0.012424888710180918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,6144,0.019239111079110038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,6144,0.331679105758667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,6144,0.010055111514197456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,5120,0.010575110713640848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,5120,0.0181342214345932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,5120,0.33493865860833066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,5120,0.009359999663299983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,7168,0.020643555455737643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,4096,0.016122667325867545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,4096,0.3386017746395535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,4096,0.008397333323955536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,3584,0.008565333154466417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,3584,0.015134221977657743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,3584,0.008005333443482717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,3584,0.33147554927402073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,3072,0.008104888929261101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,3072,0.015113777584499784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,3072,0.00775111135509279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,2560,0.007477333148320516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,2560,0.013863110707865821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,3072,0.4724524286058214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,2048,0.006008000009589725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,2560,0.3294728861914741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,2560,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,2048,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,2048,0.013538666897349887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,2048,0.32071733474731445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,1536,0.012833777401182385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,1536,0.317849768532647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,4096,0.010215111076831818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,1024,0.0041999999019834726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,1024,0.012808000048001608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,1024,0.31528178850809735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,768,0.0042053332759274375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,1024,0.006291555447710886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,768,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,1536,0.005229333208666907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,512,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,768,0.2982257737053765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,768,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,512,0.012459555433856117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,256,0.003555555724435382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,512,0.005976000179847081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,512,0.3145342138078478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,1536,0.0069413334131240845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,256,0.011778666741318174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,256,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,256,0.3274391227298313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3584,128,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3584,128,0.2968951066335042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,64,0.011769777370823754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,32,0.0035955554081334006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,32,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,65536,0.07091822226842244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,128,0.003154666680428717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3584,128,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,65536,0.11504888534545898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,16384,0.02458222210407257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3584,64,0.0031759999692440033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,65536,0.06669866376452975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,65536,0.31222134166293675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,16384,0.03463733196258545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,12288,0.02113155523935954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,16384,0.019847111569510568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,16384,0.32776888211568195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,12288,0.02883111106024848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,10240,0.0177848885456721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,12288,0.01622222198380364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,12288,0.3162311183081733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,10240,0.02576711111598545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,8192,0.01629688839117686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,10240,0.014444443914625378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,8192,0.02236266599761115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,8192,0.012056888805495368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,8192,0.3357866605122884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,7168,0.01310400002532535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,7168,0.020095111595259774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,7168,0.34867554240756565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,7168,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,6144,0.019189332922299702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,10240,0.3291066752539741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,6144,0.00998222248421775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,5120,0.010494222243626913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,6144,0.3207164340549045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,5120,0.017901332841979135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,5120,0.31491022639804417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,5120,0.009375111096435124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,4096,0.00942488925324546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,6144,0.012085333466529846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,4096,0.015853333804342482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,3584,0.008531555533409119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,4096,0.00833511104186376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,4096,0.3209137651655409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,3584,0.014779556128713818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,3072,0.007643555601437886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,3584,0.008088000118732452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,3584,0.33216532071431476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,3072,0.014478221535682678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,2560,0.006940444310506185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,3072,0.007631999750932057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,3072,0.32431289884779185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,2560,0.01407999959256914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,2048,0.005623999983072281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,2560,0.007323555648326874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,2560,0.3330266740587023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,2048,0.013461333182122974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,1536,0.0052426668504873914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,2048,0.31830400890774196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,1536,0.013245333400037555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,1536,0.006671999891599019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,1024,0.004546666724814309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,1024,0.01274222218328052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,2048,0.007003555695215861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,1024,0.3449270990159776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,768,0.0042053332759274375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,1024,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,768,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,1536,0.312882661819458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,768,0.005977777971161737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,512,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,768,0.3391831186082628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,512,0.012111110819710625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,512,0.005899555567238066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,512,0.31700534290737575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,256,0.011891555454995898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,128,0.003571555432346132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,128,0.011513777905040316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,256,0.493176884121365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,3072,128,0.3029804494645861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,64,0.0032391111469931076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,256,0.0035955554081334006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,64,0.011798222031858234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,3072,32,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,65536,0.06384266747368707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,3072,32,0.011777777638700275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,256,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,65536,0.11215021875169541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,65536,0.06559289164013334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,65536,0.34002667003207737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,16384,0.03360355562633938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,3072,128,0.005963555640644497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,12288,0.020023110840055678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,16384,0.019652444455358718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,16384,0.3248959912194146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,12288,0.028200000524520874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,10240,0.016993777619467843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,12288,0.01573955516020457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,10240,0.02545333405335744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,16384,0.02237422267595927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,10240,0.3262711101108127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,8192,0.015493333339691162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,10240,0.01348977784315745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,8192,0.021707556313938562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,8192,0.011384000380833944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,7168,0.013758222262064615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,8192,0.32995910114712185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,7168,0.020302222834693063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,7168,0.011067555182509951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,6144,0.012488889197508493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,7168,0.3318720128801134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,12288,0.3141457769605848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,6144,0.01887911061445872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,5120,0.010120000276300643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,6144,0.010011555420027839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,6144,0.3524471124013265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,5120,0.01720622181892395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,4096,0.009099555512269339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,5120,0.009336888790130615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,5120,0.3123626708984375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,4096,0.015228443675571017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,3584,0.008061333662933773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,4096,0.008292444050312042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,4096,0.34668445587158203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,3584,0.014868444866604276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,3072,0.00700711127784517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,3584,0.008026666939258575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,3584,0.32508622275458443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,3072,0.014470222923490735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,2560,0.006600889066855113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,3072,0.304928011364407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,2560,0.013796444568369122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,2560,0.007374222079912822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,2560,0.31054043769836426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,2048,0.005607111172543631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,2048,0.013471110827393003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,2048,0.34501155217488605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,2048,0.006999111009968652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,1536,0.0052435555391841466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,3072,0.007813333637184566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,1536,0.013050666285885705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,1024,0.004279110994603899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,1536,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,1536,0.33083732922871906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,1024,0.012443555725945367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,768,0.004229333251714706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,1024,0.006229333165619109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,1024,0.32603200276692706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,768,0.01236355553070704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,512,0.003968888686762916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,768,0.324710210164388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,512,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,512,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,256,0.0035111109415690103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,256,0.011765333513418833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,256,0.3462799919976129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,768,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,128,0.003240888938307762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,256,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,128,0.011397333608733283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,512,0.3225511180029975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,64,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2560,128,0.312333345413208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2560,128,0.005575110928879843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2560,32,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,65536,0.054506666130489774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,32,0.011034666664070554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,65536,0.109424889087677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,65536,0.06436266501744588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,65536,0.33779289987352157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,16384,0.03247111042340597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,16384,0.3020995457967122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,12288,0.015021334091822306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,16384,0.018963555494944256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2560,64,0.011760888828171624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,12288,0.027416000763575237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,10240,0.014078223043017916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,12288,0.015417777829700045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,12288,0.33552710215250653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,10240,0.024642666180928547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,8192,0.013421333498424955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,10240,0.012846222354306115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,10240,0.32612888018290204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,16384,0.01885688967174954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,7168,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,8192,0.02105866703722212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,8192,0.01110222190618515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,8192,0.3279955652025011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,7168,0.01944533321592543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,6144,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,7168,0.010404444403118556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,7168,0.32417689429389107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,6144,0.017832888497246634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,5120,0.00979733301533593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,6144,0.010012444522645738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,6144,0.31910043292575413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,5120,0.016502221425374348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,4096,0.008626666333940294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,5120,0.009317333499590555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,5120,0.31338487731085884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,3584,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,4096,0.00831288927131229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,4096,0.315175109439426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,3584,0.014534221755133735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,3072,0.007124444676770105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,3584,0.007965332931942409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,3584,0.3259226746029324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,3072,0.014110222458839417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,2560,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,3072,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,3072,0.3221386538611518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,2560,0.013544888959990608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,2560,0.33032798767089844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,2560,0.007342221836249034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,2048,0.005561777700980504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,4096,0.015086222026083203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,2048,0.013159111142158508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,1536,0.004927111168702443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,2048,0.007033777733643849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,2048,0.32470843527052135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,1536,0.012768000364303589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,1024,0.0042275554604000514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,1536,0.00665155549844106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,1536,0.343572457631429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,1024,0.012434666355450949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,768,0.003966222206751506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,1024,0.3313759962717692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,768,0.012166221936543783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,768,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,768,0.32258934444851345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,512,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,512,0.012186666329701742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,512,0.3323777781592475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,256,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,256,0.01180266671710544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,1024,0.006319111006127463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,256,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,128,0.0032239999208185407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,256,0.33354666497972274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,128,0.011420444481902652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,2048,128,0.30367999606662327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,64,0.002887111157178879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,2048,512,0.0058959999846087555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,2048,32,0.0028284444577164124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,65536,0.040243556102116905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,64,0.011715555356608497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,2048,32,0.011744889120260874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,65536,0.10712622271643744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,16384,0.015250666273964776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,65536,0.06258933411704169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,65536,0.3377315468258328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,16384,0.031981332434548274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,12288,0.013621333572599622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,16384,0.01758311192194621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,16384,0.3374622133043077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,12288,0.027151111099455092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,12288,0.014037332601017423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,10240,0.012053333222866058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,12288,0.34677688280741376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,10240,0.02401422295305464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,8192,0.011092444260915121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,10240,0.01275288893116845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,10240,0.3105795648362902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,8192,0.020289777053727042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,7168,0.010316444767846001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,8192,0.339919990963406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,7168,0.018565333551830716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,7168,0.010431110858917236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,7168,0.300181335873074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,6144,0.01032266683048672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,6144,0.01751377847459581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,6144,0.3334026601579454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,6144,0.010063110954231685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,5120,0.009372444616423713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,5120,0.016460445192125108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,5120,0.3303155634138319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,5120,0.009324444664849175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,4096,0.008073777788215214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,4096,0.0083137775460879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,3584,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,4096,0.3260515530904134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,3584,0.014764444695578681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,3584,0.00795377790927887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,3584,0.31823023160298664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,3072,0.006616000086069107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,8192,0.011083555718262991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,3072,0.014132445057233175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,2560,0.00564177789621883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,3072,0.007673777639865875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,3072,0.3290026717715793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,2560,0.013840888937314352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,2048,0.005244444641802046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,2560,0.007297777467303806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,4096,0.015237333046065437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,2560,0.3114986684587267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,2048,0.0131217779384719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,1536,0.004886222382386525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,2048,0.007273777491516537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,1536,0.012792000340090858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,1024,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,1536,0.31668088171217174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,1536,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,1024,0.01240711079703437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,768,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,1024,0.00629333323902554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,1024,0.31210488743252224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,768,0.012430222498046027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,512,0.003506666670242945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,768,0.00620266670982043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,768,0.3749813238779704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,512,0.3294124338361952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,2048,0.31919201215108234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,256,0.003560888684458203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,512,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,256,0.012089778151777057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,256,0.005605333381228977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,256,0.31494400236341685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,128,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,128,0.01144177797767851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,512,0.012078222301271228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1536,128,0.3162879943847656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,64,0.0029057777590221832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1536,128,0.005633777628342311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1536,32,0.00318755561278926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,65536,0.028967999749713477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,64,0.011775111158688864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1536,32,0.010710222025712332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,65536,0.10492799679438274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,65536,0.05950578053792318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,65536,0.33489688237508136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,16384,0.03154222170511881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,16384,0.016545777519543965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,12288,0.010712888505723743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,12288,0.026174222429593403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,16384,0.012860444684823355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,12288,0.31658932897779674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,10240,0.012061333490742577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,12288,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,10240,0.024088889360427856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,16384,0.3251929018232557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,10240,0.012721777790122561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,8192,0.011072888970375061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,10240,0.33957245614793563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,8192,0.01957066688272688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,8192,0.011034666664070554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,7168,0.010644444160991244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,8192,0.3036622206370036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,7168,0.018608000543382432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,7168,0.010633777413103314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,7168,0.3335164388020833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,6144,0.01758044461409251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,6144,0.00999466660949919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,6144,0.31736622916327584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,5120,0.0083137775460879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,5120,0.016192888220151264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,5120,0.33345601293775773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,5120,0.009315555294354757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,4096,0.007983110845088959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,6144,0.00903200027015474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,4096,0.01479644411140018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,3584,0.006973333656787872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,4096,0.32200800047980416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,4096,0.0083137775460879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,3584,0.014471999473041959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,3072,0.006259555617968242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,3584,0.008022222254011366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,3584,0.3107439941830105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,3072,0.013835555149449242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,2560,0.005920889063013925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,3072,0.007634667058785756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,3072,0.33499733606974286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,2560,0.013515555196338229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,2048,0.005414222263627582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,2560,0.007335999773608313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,2560,0.31423377990722656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,1536,0.004548444516128964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,2048,0.01311644415060679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,2048,0.006976889239417181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,1536,0.012802666260136498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,1536,0.006597333484225803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,1024,0.004192000048028098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,2048,0.4580862257215712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,1536,0.3251733250088162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,768,0.0038737778862317405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,1024,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,1024,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,1024,0.32058845625983345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,768,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,512,0.0035253332720862497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,768,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,768,0.3259555498758952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,512,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,256,0.0032133333798911837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,512,0.005909333212508096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,512,0.31586488087972003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,256,0.011772444678677453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,128,0.00309955556359556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,256,0.328911993238661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,128,0.011388444238238864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,64,0.0028604444944196274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,1024,128,0.28858666949801975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,128,0.00600000015563435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,1024,32,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,64,0.011378666592968835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,65536,0.024393777052561443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,1024,32,0.01180355581972334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,65536,0.10113155841827393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,16384,0.01291822228166792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,65536,0.057722667853037514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,65536,0.3262560102674696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,16384,0.03155555658870273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,1024,256,0.00564444437623024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,12288,0.011545778148704104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,16384,0.3453226619296604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,16384,0.016544888416926067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,12288,0.02529955572552151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,12288,0.013760888742076026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,10240,0.010517333116796283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,12288,0.3392008940378825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,10240,0.022314666046036616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,8192,0.009351111120647853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,10240,0.012723555167516073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,10240,0.3454613420698378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,8192,0.01930488811598884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,7168,0.008665777742862701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,8192,0.01108444482088089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,8192,0.3398088879055447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,7168,0.018246221873495314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,6144,0.007979555262459649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,7168,0.010450666977299584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,7168,0.33315467834472656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,6144,0.01724088854259915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,5120,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,6144,0.009687110781669617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,6144,0.34406667285495335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,5120,0.016163556112183463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,5120,0.009383111364311641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,5120,0.33790132734510636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,4096,0.015206222732861837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,4096,0.008287111090289222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,4096,0.3383653428819444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,3584,0.007051555646790399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,3584,0.014513777361975776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,3584,0.33161067962646484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,3072,0.006561777657932705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,4096,0.007521777517265744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,3072,0.014169777433077494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,3072,0.007675555845101674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,2560,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,3072,0.30346666442023384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,2560,0.013768889009952545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,2560,0.30948623021443683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,2048,0.0052533331844541765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,3584,0.00795911086930169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,2048,0.013092444174819522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,2048,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,2048,0.30362221929762095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,1536,0.004553777890072929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,1536,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,1536,0.3122826682196723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,1024,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,1536,0.006632888896597757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,2560,0.007366221812036302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,1024,0.012418666647540199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,768,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,1024,0.32607555389404297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,1024,0.006335999816656113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,768,0.012067555553383298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,512,0.003237333355678452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,768,0.006217777729034424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,768,0.3194968965318468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,512,0.012076444096035428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,256,0.0031893334041039148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,512,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,512,0.32423112127516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,256,0.011728000309732227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,128,0.0028853333658642242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,256,0.005703111075692707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,256,0.32269154654608834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,128,0.011413333316644033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,64,0.002879111096262932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,768,32,0.0028862222615215513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,768,128,0.29637956619262695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,768,128,0.0058906666106647914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,64,0.011368888947698804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,65536,0.01797422269980113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,768,32,0.01126488877667321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,16384,0.011104889214038849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,65536,0.056919111145867236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,65536,0.10155022144317627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,16384,0.03059200114674038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,65536,0.3287377887301975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,12288,0.009357333183288574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,16384,0.016723554995324876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,16384,0.32217333051893443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,10240,0.009330666727489894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,12288,0.32472888628641766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,10240,0.02253244486120012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,10240,0.012758221891191272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,10240,0.3221040036943224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,8192,0.00904799997806549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,8192,0.019567999574873183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,8192,0.3328862190246582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,7168,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,12288,0.024363555841975745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,7168,0.01826133330663045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,12288,0.01384888920519087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,6144,0.007328000333574083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,7168,0.010406222608354356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,7168,0.31281156010097927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,6144,0.01756622228357527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,6144,0.33667556444803876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,5120,0.007995555798212687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,6144,0.009951111343171861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,8192,0.01109333336353302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,5120,0.009352888498041365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,5120,0.37952976756625706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,4096,0.007700444095664554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,4096,0.015188443991872998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,4096,0.3352079921298557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,3584,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,3584,0.014463111758232117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,5120,0.016247999336984422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,3584,0.33749156528049046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,3584,0.008025777836640676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,3072,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,3072,0.014188443620999655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,3072,0.32963821623060435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,2560,0.0058764442801475525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,4096,0.00832622249921163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,2560,0.013735111388895245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,2560,0.3631689018673367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,2560,0.007311111523045435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,2048,0.005274666680230035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,2048,0.013119999733236102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,2048,0.3587466610802545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,1536,0.004537777768241035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,3072,0.007689778175618913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,1536,0.012865777644846173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,1536,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,1024,0.004233777936961916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,1536,0.3033573362562391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,1024,0.012530666258600024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,1024,0.31223734219868976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,1024,0.006247111078765657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,768,0.003673777812057071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,2048,0.0069919998447100324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,768,0.0124177775449223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,512,0.0032835555159383347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,768,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,768,0.31733955277336967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,512,0.012122666670216454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,512,0.006000888844331105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,512,0.3328355683220757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,256,0.011692444483439127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,256,0.005684444473849402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,128,0.0029360000044107437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,256,0.3273395432366265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,128,0.01147466649611791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,512,128,0.3002924389309353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,64,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,64,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,32,0.0028719999310043124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,512,32,0.011339555184046427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,65536,0.014814222852389017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,65536,0.09962222311231826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,65536,0.049696889188554555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,512,128,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,65536,0.4766649140252008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,512,256,0.0031884445084465873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,16384,0.029502222935358684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,12288,0.008299555215570662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,16384,0.28800177574157715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,12288,0.024653333756658766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,12288,0.013779555757840475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,12288,0.30390577846103245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,10240,0.0090844440791342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,10240,0.022452443838119507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,16384,0.009317333499590555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,10240,0.3350462118784587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,8192,0.007996444072988298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,10240,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,8192,0.019381332728597853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,7168,0.007661333514584436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,8192,0.0110880004035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,16384,0.016546666622161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,7168,0.018256000346607633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,6144,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,7168,0.3767751057942708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,6144,0.017200888858901132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,6144,0.009894222021102905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,5120,0.008029333419269985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,6144,0.35723288853963214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,5120,0.016143111719025504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,5120,0.009353777600659264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,8192,0.3254755602942573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,5120,0.3398639890882704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,4096,0.007167999943097432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,7168,0.010321777727868822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,4096,0.014811555544535318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,3584,0.0069511110583941145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,4096,0.32698488235473633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,3584,0.014494222071435718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,3072,0.0069448889957533936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,3584,0.30868445502387154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,3584,0.007689778175618913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,3072,0.014024888475735983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,2560,0.006506666541099548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,3072,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,3072,0.3069190979003906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,2560,0.013433777623706393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,2560,0.33684089448716903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,2048,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,4096,0.008282666405042013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,2048,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,2048,0.00692622239391009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,2048,0.3095831076304118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,1536,0.004616888860861461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,1536,0.012799999780125089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,1536,0.3420480092366536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,1024,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,2560,0.007297777467303806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,1024,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,1024,0.005984000033802456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,768,0.0035777779089079965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,768,0.012083555261294046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,1536,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,768,0.3418160014682346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,512,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,768,0.00620977787507905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,512,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,1024,0.3078880045149061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,512,0.32900622155931264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,256,0.0032675556010670135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,256,0.011768889096048145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,256,0.3370230992635091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,256,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,128,0.003064888839920362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,128,0.011681777735551199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,256,128,0.3086640040079753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,128,0.005620444400442972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,64,0.0028248888750871024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,256,32,0.002814222127199173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,64,0.011429333024554782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,256,512,0.005610666755172942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,65536,0.015062222878138224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,256,32,0.011382222175598145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,65536,0.04871111114819845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,65536,0.09691111246744792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,16384,0.009372444616423713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,65536,0.33713510301378036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,16384,0.028824888997607764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,12288,0.00833333366447025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,16384,0.016534222496880423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,16384,0.3240702152252197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,12288,0.02405333353413476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,10240,0.007699555820888943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,12288,0.013731555806265937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,12288,0.3252897792392307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,10240,0.02201955517133077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,8192,0.007229333122571309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,10240,0.012421333127551608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,10240,0.32050932778252494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,8192,0.018906666172875296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,8192,0.011031111081441244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,7168,0.006703111032644908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,8192,0.3343386650085449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,7168,0.018232888645595975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,7168,0.010403555300500657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,7168,0.34160621960957843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,6144,0.017218665944205392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,5120,0.006251555350091722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,6144,0.009970666633711921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,6144,0.32707733578152126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,5120,0.016336888074874878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,5120,0.008999110923873054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,5120,0.32801956600613064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,4096,0.014463111758232117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,4096,0.008247999681366814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,6144,0.00665866666369968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,3584,0.006259555617968242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,4096,0.3261351055569119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,3584,0.0144533332851198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,3584,0.008005333443482717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,3072,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,3584,0.3193964428371853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,3072,0.014144889182514615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,2560,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,3072,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,3072,0.32954666349622935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,4096,0.006588444527652528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,2048,0.006234666539563074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,2560,0.013743110828929476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,2560,0.007287999822033777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,2560,0.3204995526207818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,2048,0.013423110875818463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,1536,0.004582222137186262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,2048,0.0069724445541699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,2048,0.3226542207929823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,1536,0.012878222597969903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,1536,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,1536,0.32049422793918186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,1024,0.012451555165979596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,768,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,1024,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,1024,0.3033591111501058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,768,0.012416889270146688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,512,0.003216000066863166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,768,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,768,0.31939734352959526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,512,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,512,0.32989954948425293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,512,0.005909333212508096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,256,0.0029057777590221832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,1024,0.00388355553150177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,256,0.01182666669289271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,128,0.0028328889360030494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,256,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,256,0.3186355431874593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,128,0.011558222273985544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,64,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,128,128,128,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,128,32,0.0025173332542181015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,128,128,128,0.2991528775956896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,65536,0.013537777794731988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,64,0.011445333560307821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,128,32,0.011130666567219628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,65536,0.09652888774871826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,16384,0.02920266654756334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,12288,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,10240,0.007344000041484833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,12288,0.024519999821980793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,8192,0.0069146665434042616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,10240,0.02236533330546485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,7168,0.006728888799746831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,8192,0.019318222999572754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,6144,0.006625777731339137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,7168,0.018273777431911893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,5120,0.006072000082996156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,6144,0.017136888371573556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,16384,0.00832177781396442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,4096,0.005967999911970562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,5120,0.016132444143295288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,3584,0.006214222146405115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,3072,0.006030222194062338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,4096,0.015117333994971381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,3584,0.014532445205582513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,2560,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,2048,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,2560,0.013524444566832649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,1536,0.004530666602982415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,2048,0.013454222016864352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,1024,0.0038613333470291565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,1536,0.012825777961148156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,768,0.003581333491537306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,768,0.012111110819710625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,512,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,512,0.012135110795497894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,256,0.0029084444459941653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,3072,0.01387288918097814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,128,0.0029582221888833572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,256,0.011719110939237805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,64,0.002918222298224767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,128,0.01144977741771274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,64,32,0.0029057777590221832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,64,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,65536,0.013447110851605734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,32,0.011076444553004371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,16384,0.008083555433485243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,65536,0.09515733189053005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,12288,0.007679999702506595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,16384,0.02916533417171902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,10240,0.007381333245171442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,12288,0.024608888559871252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,8192,0.0075137780772315125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,10240,0.022359111242824133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,7168,0.006911110960774952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,8192,0.019384000036451552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,6144,0.006642666541867786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,64,1024,0.012464000119103326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,5120,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,5120,0.01644711030854119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,4096,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,4096,0.014871110518773397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,3584,0.00638133328821924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,3584,0.014480888843536377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,7168,0.01829600003030565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,3072,0.006257777826653586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,2560,0.0063680000603199005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,2048,0.006175111151403851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,6144,0.017242666747834947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,2560,0.013677332964208392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,1536,0.00461422238085005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,1024,0.003826666623353958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,2048,0.012851555314328937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,1536,0.012823111481136747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,768,0.003576000117593341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,1024,0.012271111210187277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,512,0.003228444399105178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,256,0.0029475556479560006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,768,0.012090666426552666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,128,0.0028568889117903183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,256,0.01145600030819575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,128,0.011446221835083432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,64,0.0028977776981062363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,128,32,32,0.002887111157178879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,64,0.011382222175598145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,32,0.010378667049937779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,3072,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,16384,0.35368887583414715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,16384,0.21397956212361655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,128,32,512,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,12288,0.26798399289449054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,16384,0.14289955298105875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,12288,0.16621600257025823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,16384,0.3405315611097548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,12288,0.11091199848386978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,10240,0.2051173316107856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,10240,0.1352577739291721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,12288,0.378565337922838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,8192,0.16815733909606934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,10240,0.09595111343595718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,8192,0.11774222056070964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,10240,0.3512008984883626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,8192,0.07509333557552762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,7168,0.1469822194841173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,8192,0.3441013230217828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,7168,0.0989982220861647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,6144,0.12908977932400173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,7168,0.07067288955052693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,6144,0.0867404474152459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,7168,0.32808444235059947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,6144,0.05616177452935112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,5120,0.1074862215254042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,6144,0.3303626643286811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,5120,0.07474666833877563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,4096,0.08849333392249213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,5120,0.05284533235761854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,4096,0.06846133205625746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,5120,0.3322426742977566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,3584,0.07792977492014568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,4096,0.04201511210865445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,4096,0.32865956094529897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,3584,0.05739733245637682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,3072,0.0670035547680325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,3584,0.03601333167817857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,3072,0.05086044470469157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,3584,0.3289857705434163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,2560,0.057780444622039795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,3072,0.034232000509897866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,3072,0.3321315447489421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,2560,0.04487733377350701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,2048,0.049048887358771436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,2560,0.028840889533360798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,2048,0.03927733169661628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,2560,0.34235554271274143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,1536,0.03639199998643663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,2048,0.02387111054526435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,2048,0.3077262242635091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,1536,0.03329955538113912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,1024,0.025938666529125635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,1536,0.019375110665957134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,1536,0.32702933417426217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,1024,0.028600000672870215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,768,0.020655999581019085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,1024,0.015145777000321282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,1024,0.31354933314853245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,768,0.023737778266270954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,512,0.014319999350441827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,768,0.013471110827393003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,768,0.2952382299635145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,256,0.008985777695973715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,512,0.021951110826598272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,512,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,512,0.30813421143425834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,256,0.018534221582942538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,256,0.010455999937322406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,128,0.007963555554548899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,256,0.2889875570933024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,64,0.006934222247865465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,128,0.01741333305835724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,65536,32,0.006945778098371293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,64,0.017924444542990792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,65536,128,0.32044267654418945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,65536,32,0.021517333057191636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,65536,0.324689785639445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,65536,0.2563262250688341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,16384,0.08846755822499593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,16384,0.076200889216529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,65536,128,0.010732444624106089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,65536,0.16545155313279894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,12288,0.06813600328233507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,16384,0.04621155394448174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,12288,0.05954755677117241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,16384,0.33724533187018496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,10240,0.05911022424697876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,12288,0.03613600134849548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,65536,0.3104417853885227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,12288,0.31785689459906685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,10240,0.05056622293260363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,8192,0.046421332491768726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,10240,0.03591377867592706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,8192,0.04478666517469618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,10240,0.32672354910108775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,7168,0.0421013335386912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,8192,0.02714666724205017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,8192,0.3126968807644314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,6144,0.038555555873446994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,7168,0.024674667252434626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,6144,0.03506133291456435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,7168,0.33266666200425893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,6144,0.020625778370433383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,6144,0.32351022296481663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,5120,0.03155111273129781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,5120,0.03144355614980062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,5120,0.02082577844460805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,7168,0.039507554637061224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,4096,0.025324443976084392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,4096,0.02905155552758111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,4096,0.314376884036594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,3584,0.023388443721665278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,4096,0.015628443823920358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,3584,0.024533333049880132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,3584,0.33005865414937335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,3072,0.02036533256371816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,3584,0.015222221612930298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,5120,0.33756976657443577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,3072,0.022869333624839783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,2560,0.01757244434621599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,3072,0.34043288230895996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,2560,0.0210853334930208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,2560,0.012282667060693106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,2560,0.3270648850335015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,2048,0.015671999918089975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,2048,0.01902933253182305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,2048,0.33496801058451336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,1536,0.011657777759763928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,3072,0.013775111072593264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,1536,0.017966222431924608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,1536,0.009335999687512716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,1024,0.00943288869327969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,1536,0.31498222880893284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,1024,0.016369778248998854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,1024,0.31517065895928276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,1024,0.00796977761718962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,768,0.008001777860853408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,2048,0.011337777806652917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,768,0.014848889576064216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,512,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,768,0.31898755497402614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,768,0.007311999797821045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,512,0.014450665977266101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,256,0.004512000001139111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,512,0.00702400008837382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,512,0.3101422256893582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,256,0.013047110703256396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,128,0.0041982221106688184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,256,0.006657777975002925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,256,0.3322960005866157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,128,0.012816000315878125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,64,0.003560888684458203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,16384,128,0.00666222224632899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,16384,128,0.29764265484280056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,16384,32,0.0036159998012913596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,64,0.01311822235584259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,16384,32,0.014135999812020196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,65536,0.2602168983883328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,65536,0.20691288842095268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,16384,0.06912889083226521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,65536,0.1563075515958998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,16384,0.06279555294248793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,12288,0.053842667076322764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,16384,0.04158044523662991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,12288,0.04938222302330864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,16384,0.4510969056023492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,12288,0.3178328937954373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,65536,0.34112977981567383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,10240,0.046943998999065824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,10240,0.04202311237653097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,10240,0.02815199891726176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,10240,0.2920284536149767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,8192,0.03810844487614102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,8192,0.3112151092953152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,7168,0.03356444504525926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,12288,0.03152622116936578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,7168,0.033414221472210355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,7168,0.28871112399631077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,8192,0.038344889879226685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,6144,0.02959377898110284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,6144,0.029651555750105116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,6144,0.3156977759467231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,5120,0.02544622288809882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,6144,0.018710222509172227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,5120,0.027354665928416785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,5120,0.3341395590040419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,5120,0.016588444511095684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,4096,0.021498666869269475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,7168,0.020415110720528495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,4096,0.024400000770886738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,3584,0.018985778093338013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,4096,0.014639111028777229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,4096,0.3189920054541694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,3584,0.021731555461883545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,3584,0.01367022262679206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,3072,0.01664622293578254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,8192,0.023618666662110224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,3584,0.31851911544799805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,2560,0.014177777700954013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,3072,0.020257777637905545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,3072,0.012724444270133972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,3072,0.3262222343020969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,2560,0.019323555959595572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,2048,0.01332800007528729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,2560,0.011649777491887411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,2048,0.017835555805100333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,2048,0.009953777823183272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,2048,0.33695377243889707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,1536,0.010219555762079027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,1536,0.01688799924320645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,1536,0.327218665017022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,1536,0.008691555923885768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,1024,0.008047999607192146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,1024,0.0151555554734336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,2560,0.3254275586869982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,768,0.00721777809990777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,1024,0.32222043143378365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,1024,0.007690666450394525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,768,0.014098667436175875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,512,0.005250666704442766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,768,0.007357333269384172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,768,0.3252737787034777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,512,0.013428444663683573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,256,0.0041991112132867174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,512,0.3336319923400879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,256,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,256,0.006573333508438534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,256,0.32891911930508083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,128,0.003855111284388436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,128,0.012740444805887012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,12288,128,0.28469599617852104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,64,0.0035182221068276297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,128,0.006632888896597757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,12288,32,0.0034755555291970572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,64,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,12288,32,0.013464888764752282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,65536,0.21336355474260119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,65536,0.20012177361382377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,65536,0.1373466650644938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,16384,0.058648890919155545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,12288,512,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,16384,0.06137333313624064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,65536,0.3115031189388699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,12288,0.04501422246297201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,16384,0.03884444302982754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,16384,0.3041120105319553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,12288,0.04749066630999247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,12288,0.030384888251622517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,10240,0.03819288810094198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,12288,0.333604441748725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,10240,0.04125066598256429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,8192,0.03163644340303209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,10240,0.026543999711672466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,10240,0.32496621873643666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,7168,0.02797599964671665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,8192,0.03680355681313409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,8192,0.022081777453422546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,7168,0.03208533260557387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,8192,0.3326142099168566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,6144,0.025082666012975905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,7168,0.020243555307388306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,7168,0.32399021254645455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,5120,0.02146400014559428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,6144,0.018283555905024212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,6144,0.029104888439178467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,6144,0.3321617709265815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,5120,0.025095999240875244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,4096,0.01829955478509267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,5120,0.016009777784347534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,5120,0.319722678926256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,4096,0.024031110935741003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,3584,0.01575911045074463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,4096,0.3177848921881782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,3584,0.021083555287784998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,3072,0.013714666995737286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,3584,0.012924444344308643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,3584,0.3179386721716987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,3072,0.020240889655219186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,2560,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,3072,0.31604356235928005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,2560,0.01887822151184082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,2560,0.01071466671095954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,2048,0.011744889120260874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,4096,0.013861333330472311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,2048,0.017610667480362784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,2048,0.009266667068004608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,2048,0.33284976747300893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,1536,0.008784888519181145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,3072,0.012080888781282636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,1536,0.016187555260128446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,1024,0.007368000017272101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,1536,0.3260684543185764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,1024,0.014512888259357877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,2560,0.3087315559387207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,1024,0.007459555235173967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,768,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,1024,0.33402132987976074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,768,0.013825777504179211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,768,0.007260444263617198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,512,0.004925333377387789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,768,0.3117751015557183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,512,0.013461333182122974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,512,0.31322399775187176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,512,0.006960888703664144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,256,0.0041884444653987885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,1536,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,256,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,128,0.003593777616818746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,256,0.006227555374304454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,256,0.2997724480099148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,128,0.012446222205956778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,10240,128,0.00664444433318244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,64,0.0035128887328836652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,10240,32,0.0038897775941424896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,10240,128,0.2830604447258843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,64,0.012728888955381183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,10240,32,0.013188444077968597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,65536,0.1654693285624186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,65536,0.15886133246951634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,16384,0.04527466826968723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,65536,0.07791111204359266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,16384,0.05002133382691277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,16384,0.31304976675245494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,65536,0.33659911155700684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,12288,0.04012622104750739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,12288,0.32666132185194224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,12288,0.0188400000333786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,16384,0.025118221839269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,10240,0.03130933311250474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,12288,0.03629511263635423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,10240,0.03404355711407132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,10240,0.016688888271649677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,8192,0.025295999315049913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,10240,0.3328835434383816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,8192,0.029776000314288672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,7168,0.022875555687480505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,8192,0.015557333827018738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,8192,0.3449057737986247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,7168,0.02643377747800615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,6144,0.020815110868877836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,7168,0.013474666410022311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,7168,0.3367484410603841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,6144,0.024305777417288885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,5120,0.01754755609565311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,6144,0.01240888900227017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,6144,0.3324551052517361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,5120,0.02164711058139801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,4096,0.014745778507656522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,5120,0.011423110961914062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,5120,0.3369057708316379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,4096,0.020599999361568026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,3584,0.013160000244776407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,4096,0.009840000006887648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,4096,0.3424453205532498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,3584,0.018971555762820773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,3072,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,3584,0.009199110998047723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,3584,0.32330576578776044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,3072,0.01793955597612593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,2560,0.01071111112833023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,3072,0.008662222160233391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,3072,0.32718578974405926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,2560,0.01719911065366533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,2048,0.009832888841629028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,2560,0.008229333493444655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,2048,0.015821332732836407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,1536,0.007890666524569193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,2048,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,2560,0.46718846427069766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,1536,0.014112888111008538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,1024,0.005963555640644497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,1536,0.006978666616810693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,1536,0.3203715483347575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,1024,0.013751111096805997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,768,0.00498933345079422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,1024,0.006585777633719974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,768,0.01348444405529234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,1024,0.4019928773244222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,512,0.003967111309369405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,768,0.3241848945617676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,768,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,512,0.013050666285885705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,256,0.003576888806290097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,2048,0.32294223043653697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,512,0.31674665874905056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,256,0.012108444339699216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,128,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,256,0.00591822216908137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,256,0.3273724450005425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,128,0.012146666646003723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,64,0.0032355555643637977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,128,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,8192,128,0.28917956352233887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,64,0.012865777644846173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,8192,32,0.01276266657643848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,65536,0.15199289056989881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,8192,512,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,65536,0.15459733539157444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,16384,0.042502221133973866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,65536,0.0755866699748569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,16384,0.048835555712381996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,8192,32,0.0032355555643637977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,16384,0.2817111015319824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,65536,0.3582879967159695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,12288,0.03401866555213928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,16384,0.0229777776532703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,12288,0.03837688763936361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,12288,0.017959111266665988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,10240,0.0292871097723643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,12288,0.3102533287472195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,10240,0.03294577863481309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,10240,0.016159999701711867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,8192,0.023642667465739783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,10240,0.3494817680782742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,8192,0.029667556285858154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,7168,0.021352888809310064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,8192,0.014838222000333997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,8192,0.3394666777716742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,7168,0.025797333982255723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,6144,0.019700444406933255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,7168,0.3515048821767171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,6144,0.023372444841596816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,6144,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,6144,0.33675376574198407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,5120,0.016380444169044495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,5120,0.021943999661339655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,5120,0.3420400089687771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,4096,0.014911999305089315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,5120,0.011121778024567498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,7168,0.01311555587583118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,4096,0.020635555187861126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,3584,0.012261333564917246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,4096,0.009309333231714036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,4096,0.32844532860649955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,3584,0.018926221463415358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,3072,0.011063111325105032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,3584,0.00869333330127928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,3584,0.31953689787122935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,3072,0.017901332841979135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,2560,0.009955555200576782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,3072,0.00831911133395301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,3072,0.3259955512152778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,2560,0.01683555543422699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,2048,0.009706666900051964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,2560,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,2560,0.3306559986538357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,2048,0.015192889504962497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,1536,0.0074746666683091064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,2048,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,2048,0.3307502269744873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,1536,0.014130666851997375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,1024,0.005982222242487802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,1536,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,1536,0.3218364450666639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,1024,0.013795555465751223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,768,0.004884444591071871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,1024,0.006305777778228124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,1024,0.30906666649712455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,768,0.01311555587583118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,512,0.004554666578769684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,768,0.334640900293986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,512,0.012448888685968188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,512,0.005975111077229182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,512,0.32255556848314076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,256,0.00423555572827657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,256,0.012073777616024017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,256,0.33632710244920516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,256,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,128,0.003605333467324575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,768,0.006197333335876465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,7168,128,0.29487111833360463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,64,0.003195555466744635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,7168,128,0.005920889063013925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,7168,32,0.003495111233658261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,64,0.012096000214417776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,65536,0.12515466743045384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,32,0.012526222401195101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,65536,0.14824088414510092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,16384,0.039270222187042236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,65536,0.0743537810113695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,16384,0.047316445244683154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,7168,128,0.01204177737236023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,12288,0.03369244601991441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,65536,0.3364204565684001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,16384,0.418762657377455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,12288,0.310153775744968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,12288,0.017613333132531907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,10240,0.028246220615175035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,10240,0.031503111124038696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,10240,0.3211795488993327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,10240,0.015829333000712924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,8192,0.021896888812383015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,16384,0.021905778182877436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,8192,0.02847555610868666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,7168,0.019370666808552213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,12288,0.03715644280115763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,8192,0.3199991120232476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,8192,0.014024888475735983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,6144,0.017296888762050204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,7168,0.025030222203996446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,7168,0.012889777620633444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,6144,0.023017777336968318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,7168,0.33534222178988987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,5120,0.014870221416155497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,6144,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,5120,0.02130844526820713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,5120,0.01051111105415556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,5120,0.33238222863939076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,4096,0.013462222284740873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,4096,0.019985778464211356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,4096,0.3459182315402561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,4096,0.009017777939637503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,3584,0.011858666936556498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,3584,0.018561777141359117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,6144,0.34901245435078937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,3072,0.010423999693658618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,3584,0.3434986538357205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,3584,0.008353778057628209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,3072,0.017954667409261067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,3072,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,2560,0.009369778136412302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,3072,0.33262133598327637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,2560,0.0158933334880405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,2048,0.008989333278603023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,2560,0.007631111476156447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,2560,0.3315884537167019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,2048,0.014748444159825643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,1536,0.006802666518423293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,2048,0.007282666862010956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,2048,0.3258879979451497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,1536,0.013832000394662222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,1024,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,1536,0.0069146665434042616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,1536,0.30217332310146755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,1024,0.013442666994200813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,768,0.004554666578769684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,1024,0.3354862266116672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,768,0.013136000268989138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,768,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,768,0.3092213206821018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,512,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,512,0.3313528961605496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,1024,0.006330666856633292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,256,0.003583999971548716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,512,0.005968889014588461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,256,0.012111999922328524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,512,0.0038844446341196695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,256,0.005609777652555042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,128,0.003228444399105178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,256,0.3421324359046088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,128,0.011919111013412476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,6144,128,0.005973333285914526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,6144,128,0.3098355664147271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,32,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,64,0.012096000214417776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,65536,0.1086568898624844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,6144,32,0.011705777711338468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,65536,0.14454844262864855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,16384,0.035275555319256253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,65536,0.07096978028615315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,6144,64,0.003173333282272021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,65536,0.3271911144256592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,16384,0.04595999916394552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,12288,0.028788444068696763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,16384,0.021127111381954614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,12288,0.03524000114864773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,16384,0.339591105779012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,12288,0.016892444756295946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,10240,0.0266266663869222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,12288,0.3261715571085612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,8192,0.018466666340827942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,10240,0.03159733282195197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,10240,0.015286222100257874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,10240,0.3152560128106011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,8192,0.02734400000837114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,7168,0.01640977793269687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,8192,0.013822221921549903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,8192,0.32858578364054364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,7168,0.024519110719362896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,6144,0.015244444211324057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,7168,0.012372444073359171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,7168,0.321361780166626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,6144,0.02260977857642704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,5120,0.012746666868527731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,6144,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,6144,0.33240000406901044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,5120,0.020837333467271592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,4096,0.011350221931934357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,5120,0.00979200005531311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,5120,0.30819911426968044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,3584,0.011196444431940714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,4096,0.01990133358372582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,4096,0.00852888905339771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,4096,0.32409776581658256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,3072,0.009988444546858469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,3584,0.018253333038753934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,3584,0.3269466559092204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,3072,0.016862221890025668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,2560,0.008942222429646386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,3072,0.2999182277255588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,3072,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,2560,0.015836444165971544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,2560,0.30692889955308705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,2560,0.007311111523045435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,2048,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,2048,0.01481244464715322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,2048,0.31377243995666504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,3584,0.008016000191370646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,2048,0.006990222467316522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,1536,0.01386577801571952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,1536,0.32722311549716526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,1024,0.005256888767083486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,1024,0.013463111387358772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,1024,0.3281075424618191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,1536,0.00665866666369968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,768,0.004572444491916233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,1024,0.0063031112982167145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,768,0.012796444197495779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,1536,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,768,0.3408079942067464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,768,0.006212444355090459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,512,0.01279022213485506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,512,0.33075910144382054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,512,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,256,0.003588444242874781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,256,0.012120000190205045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,256,0.33361421691046816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,256,0.005663110978073544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,128,0.0032044444233179092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,128,0.011953777737087674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,512,0.004284444368547863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,64,0.0028880000528362063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,5120,128,0.2999937799241808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,5120,32,0.00319377767542998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,64,0.012510221865442065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,5120,32,0.012127111355463663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,65536,0.08951377868652344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,65536,0.10690755314297146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,65536,0.06863200002246432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,16384,0.028833776712417603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,5120,128,0.005642666584915585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,16384,0.034796443250444196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,16384,0.020289777053727042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,12288,0.025495999389224585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,16384,0.3131004439459907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,12288,0.02782400117980109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,12288,0.2960408793555366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,12288,0.01684177749686771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,65536,0.3391466670566135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,10240,0.024060444699393377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,10240,0.01512888901763492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,10240,0.32871821191575795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,8192,0.01628177695804172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,8192,0.022036444809701707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,8192,0.3425617747836643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,7168,0.014134221606784396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,8192,0.013019555144839816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,10240,0.022051556242836848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,6144,0.013280889226330651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,7168,0.3231360117594401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,7168,0.01180355581972334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,6144,0.019000000423855253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,5120,0.011342222491900126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,6144,0.01051111105415556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,6144,0.3182035552130805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,5120,0.01791555517249637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,4096,0.01038755559258991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,5120,0.009776888622177972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,4096,0.016021332807011075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,5120,0.4184266726175944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,4096,0.008349333372380998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,4096,0.28013955222235787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,7168,0.02032088902261522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,3584,0.015190222197108798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,3072,0.008404444489214156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,3584,0.00796977761718962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,3584,0.30369332101609975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,3072,0.014886221951908536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,3072,0.007698666718271043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,3072,0.31421155399746364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,2560,0.014084445105658637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,2560,0.007374222079912822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,2560,0.32180799378289116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,2048,0.007483555210961237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,3584,0.009032888544930352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,2048,0.013502221968438892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,1536,0.004937777916590373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,2048,0.007050666544172499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,2048,0.3279457887013753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,1536,0.013166222307417126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,1024,0.0042053332759274375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,1536,0.00666222224632899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,2560,0.007639110916190677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,1536,0.3572675651974148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,1024,0.01274400038851632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,768,0.00423555572827657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,1024,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,1024,0.32240088780721027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,768,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,512,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,768,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,768,0.3350871139102512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,512,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,256,0.003552000141806073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,512,0.32446667883131236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,512,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,256,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,128,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,256,0.005671999934646819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,256,0.32492976718478733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,128,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,64,0.003007111036115222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,4096,128,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,4096,128,0.3150257799360487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,4096,32,0.0032382222513357797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,64,0.011756444142924415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,65536,0.07771999968422784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,4096,32,0.011395555403497485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,65536,0.10523644420835708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,16384,0.02688088847531213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,65536,0.06715022193060981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,16384,0.03336266676584879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,65536,0.3408391210767958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,12288,0.02237155536810557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,16384,0.020294222566816542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,16384,0.33812443415323895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,12288,0.02628622286849552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,12288,0.016417778200573392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,12288,0.3407013416290283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,10240,0.02334755493534936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,10240,0.014839111102951897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,10240,0.33889955944485134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,8192,0.016881777180565726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,8192,0.0214008887608846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,8192,0.3428142335679796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,7168,0.0158942225906584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,10240,0.01867911054028405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,7168,0.019653333557976615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,7168,0.011318222516112857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,6144,0.012683555483818054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,6144,0.018567999203999836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,8192,0.012490666574902005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,6144,0.34116533067491317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,5120,0.010812444819344414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,6144,0.010030222435792288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,5120,0.01721333298418257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,7168,0.34439555803934735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,4096,0.009666666388511658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,5120,0.33374399609035915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,5120,0.009376889301670922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,4096,0.016011555989583332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,3584,0.008470222353935242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,4096,0.00830844458606508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,4096,0.32800889015197754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,3584,0.015446222490734525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,3072,0.00816533300611708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,3584,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,3584,0.3301466570960151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,3072,0.014479110638300577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,2560,0.007224000162548489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,3072,0.32703910933600533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,2560,0.013838222457302941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,2560,0.007304000357786815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,2048,0.0059368887709246735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,2048,0.013432888521088494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,2048,0.3486435678270128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,3072,0.007994666695594788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,1536,0.005128889034191768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,2048,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,1536,0.012820444173283048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,1024,0.004483555754025777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,1536,0.006680000159475539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,1536,0.3303537898593479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,2560,0.3190622329711914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,768,0.0038693332009845306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,1024,0.006304889089531369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,1024,0.32469243473476833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,768,0.012465777496496836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,512,0.0038773334688610504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,768,0.006022222340106964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,768,0.32564711570739746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,512,0.012416000167528788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,256,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,512,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,512,0.3332035541534424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,256,0.011776888536082374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,128,0.003216888962520493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,256,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,256,0.32955288887023926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,1024,0.012416889270146688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,64,0.003208000005947219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3584,128,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3584,32,0.0031591111587153543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,64,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,65536,0.06814489099714491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,65536,0.10209333234363133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,128,0.011810666984981961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,65536,0.3390977647569444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3584,128,0.3076808982425266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,65536,0.06651466422610812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,16384,0.02328355610370636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3584,32,0.011742221812407175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,16384,0.03184533450338576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,12288,0.020837333467271592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,16384,0.3332684569888645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,12288,0.025677333275477093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,12288,0.3291342258453369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,12288,0.016201777590645682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,10240,0.017249777913093567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,10240,0.022817777262793645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,10240,0.32171376546223956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,10240,0.014473777678277759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,8192,0.014505777094099255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,8192,0.02095555596881443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,16384,0.019670221540662978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,7168,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,8192,0.3302968872918023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,8192,0.011505777637163797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,7168,0.019588443968031142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,6144,0.01185955521133211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,7168,0.010514666636784872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,7168,0.3236337767706977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,6144,0.01816177699300978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,5120,0.010249777800507015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,6144,0.0103502223889033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,6144,0.3426728778415256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,5120,0.016925333274735343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,4096,0.00960622231165568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,5120,0.009340444372759925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,5120,0.33240267965528697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,4096,0.015132443772421943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,3584,0.00850311087237464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,4096,0.008310221963458592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,4096,0.3434026771121555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,3584,0.014523555835088095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,3072,0.007696000238259633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,3584,0.007937777373525832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,3584,0.338135110007392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,3072,0.014110222458839417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,2560,0.006969778074158563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,3072,0.007617777420414819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,3072,0.3252195517222087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,2560,0.013817778064144982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,2048,0.006278222219811545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,2560,0.007415999968846639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,2560,0.33771289719475645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,2048,0.013481777575280933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,1536,0.0053022222386466126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,2048,0.0069857777820693115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,2048,0.3330755498674181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,1536,0.012960889273219638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,1024,0.0045422220395671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,1536,0.006800000038411882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,1536,0.32463023397657603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,1024,0.012704888979593912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,768,0.004230222354332606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,1024,0.006234666539563074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,1024,0.3275609016418457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,768,0.012359110845459832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,512,0.0036551112102137674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,768,0.3307644526163737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,768,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,512,0.012130666938092975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,256,0.003506666670242945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,512,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,512,0.32717421319749623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,256,0.011760888828171624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,128,0.0032524443748924467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,256,0.005616888817813661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,256,0.3366764386494954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,128,0.011586666935020022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,64,0.003195555466744635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,3072,128,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,3072,32,0.0029297777348094513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,3072,128,0.3080328835381402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,64,0.012057777908113269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,65536,0.059567113717397056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,3072,32,0.011622221933470832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,65536,0.09962222311231826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,16384,0.02106933295726776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,65536,0.06482577986187406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,16384,0.030913776821560327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,65536,0.3320355680253771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,12288,0.01891288823551602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,16384,0.01944977707333035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,16384,0.3403955565558539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,12288,0.025269332859251235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,10240,0.016093333562215168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,12288,0.01574577722284529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,10240,0.02285066743691762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,12288,0.3333173327975803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,8192,0.014529777897728814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,10240,0.01348000019788742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,8192,0.020637333393096924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,8192,0.011394666300879585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,10240,0.45944801966349286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,8192,0.3222746584150526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,7168,0.019313777486483257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,6144,0.011614222493436603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,7168,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,7168,0.40417154630025226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,6144,0.017599110802014668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,6144,0.331371545791626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,6144,0.010029333333174387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,5120,0.016175111134847004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,7168,0.012763555679056378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,5120,0.009351111120647853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,5120,0.3448799981011285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,4096,0.009123555488056606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,4096,0.015374221735530429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,4096,0.3510177665286594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,4096,0.008384000096056197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,5120,0.010256888965765635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,3584,0.008307555483447181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,3584,0.01460000044769711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,3584,0.007998222278224098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,3072,0.007541333635648091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,3584,0.310153775744968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,3072,0.014134221606784396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,2560,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,3072,0.007696000238259633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,3072,0.3464026716020372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,2560,0.013777777552604675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,2048,0.005987555616431766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,2560,0.007345777418878343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,2560,0.32763200336032444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,2048,0.013417777915795645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,1536,0.004990222139490975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,2048,0.0069431112044387394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,2048,0.33759021759033203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,1536,0.013070222404268054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,1024,0.00425155543618732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,1536,0.006664888726340399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,1536,0.3214586575826009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,1024,0.012498666842778524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,768,0.003988444391224119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,1024,0.006239111224810283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,1024,0.3317955599890815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,768,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,512,0.003535110917356279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,768,0.006019555446174409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,768,0.3184995651245117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,512,0.012108444339699216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,256,0.0032266666077905228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,512,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,512,0.3424035443200006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,256,0.011787555283970304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,128,0.0031386667655573953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,256,0.005888000130653381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,256,0.32163111368815106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,128,0.011581333147154914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,64,0.0029333333174387612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2560,128,0.005578666511509154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2560,32,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2560,128,0.31062399016486275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,64,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,65536,0.04911555515395271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2560,32,0.011753777662913004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,65536,0.09554844432406956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,16384,0.01906044450071123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,65536,0.06349955664740668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,65536,0.3314693239000109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,16384,0.030250665214326646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,12288,0.016557332542207506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,16384,0.3421928882598877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,16384,0.019256000717480976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,12288,0.0247013337082333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,12288,0.014827556080288358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,12288,0.33088000615437824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,10240,0.022475555539131165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,10240,0.013013333082199097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,10240,0.32243643866644967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,8192,0.01311822235584259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,8192,0.020590222544140287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,8192,0.011027555498811932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,8192,0.3435431056552463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,7168,0.011777777638700275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,10240,0.014508444401952954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,7168,0.018595554762416415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,7168,0.01050222251150343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,6144,0.017256889078352187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,6144,0.009723555710580613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,5120,0.0123848890264829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,6144,0.44371912214491105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,5120,0.01592088904645708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,5120,0.009325332939624786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,4096,0.010026666853162978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,5120,0.39609599113464355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,6144,0.010811555716726514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,4096,0.014767999450365702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,7168,0.32462933328416615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,4096,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,4096,0.35955556233723956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,3584,0.014473777678277759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,3072,0.008433777425024245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,3584,0.33417333496941465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,3584,0.008029333419269985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,3072,0.014119999276267158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,2560,0.007342221836249034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,3072,0.007675555845101674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,3072,0.3214755588107639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,2560,0.013815999858909182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,2560,0.007335999773608313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,2560,0.35306310653686523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,2048,0.006276444428496891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,3584,0.009381333159075843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,2048,0.01308444473478529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,1536,0.004987555659479565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,2048,0.32956711451212567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,1536,0.01277599980433782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,1536,0.29774755901760525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,1536,0.006603555546866522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,1024,0.0042053332759274375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,1024,0.01239911135700014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,1024,0.3104746606614855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,768,0.0038951109680864546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,1024,0.006312888943486744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,2048,0.007010666860474481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,768,0.012127999630239276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,512,0.0037297776175869834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,768,0.3374142116970486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,768,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,512,0.012068444656001197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,256,0.003519111209445529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,512,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,512,0.3280506663852268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,256,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,128,0.0032462223122517266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,256,0.005620444400442972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,128,0.011719110939237805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,2048,128,0.005621333503060871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,128,0.3041244347890218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,64,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,2048,32,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,64,0.011339555184046427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,2048,32,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,65536,0.03967377874586318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,65536,0.09524977869457668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,2048,256,0.3140453232659234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,65536,0.3365351094139947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,65536,0.061936000982920326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,16384,0.014760888285107084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,16384,0.02979644470744663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,16384,0.018197332819302876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,16384,0.3240657647450765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,12288,0.02439911166826884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,12288,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,12288,0.322035551071167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,10240,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,10240,0.02203111184967889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,10240,0.012446222205956778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,10240,0.3349600103166368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,8192,0.012323555847009024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,12288,0.013454222016864352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,8192,0.01891288823551602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,7168,0.011085333095656501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,8192,0.011075555450386472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,8192,0.3176035616132948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,7168,0.01791111131509145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,6144,0.010472888747851053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,7168,0.0103493332862854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,7168,0.33151645130581325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,6144,0.01644622286160787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,5120,0.009718221922715506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,6144,0.31724800003899467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,6144,0.00999466660949919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,5120,0.01555288831392924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,5120,0.009264888862768808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,4096,0.00889955543809467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,5120,0.3073235617743598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,3584,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,4096,0.008346666892369589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,4096,0.014799111419253878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,4096,0.32431644863552517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,3584,0.01481244464715322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,3072,0.006530666516886816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,3584,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,3584,0.3367191155751546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,3072,0.014092443717850579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,2560,0.005965333431959152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,3072,0.007633777956167857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,3072,0.32048267788357204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,2560,0.013776889277829064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,2048,0.0053093334039052325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,2560,0.007376000285148621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,2560,0.33164000511169434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,2048,0.013090666797426013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,1536,0.004896000027656555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,2048,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,2048,0.3155031204223633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,1536,0.012726222475369772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,1024,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,1536,0.3332497808668348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,1024,0.012443555725945367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,768,0.003640000190999773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,1024,0.31401154730055064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,1024,0.006223111102978389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,768,0.01239822225438224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,512,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,768,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,768,0.31082044707404244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,512,0.012121777567598553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,512,0.0058782220714622075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,512,0.34563999705844456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,256,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,1536,0.006683555742104848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,256,0.011752888560295105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,128,0.003149333306484752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,256,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,256,0.3363040023379856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,128,0.011700444751315646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,64,0.0029066666546795103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1536,128,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1536,32,0.002833777831660377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1536,128,0.31564444965786403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,64,0.011494222614500256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,65536,0.029042667812771265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1536,32,0.011727111207114326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,65536,0.091848889986674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,65536,0.058079110251532666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,16384,0.013167999684810638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,65536,0.3435048792097304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,16384,0.02958844436539544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,12288,0.012083555261294046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,16384,0.016336000627941556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,16384,0.3548115624321832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,12288,0.023399111297395494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,10240,0.011146666275130378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,12288,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,12288,0.3368399937947591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,10240,0.021295999487241108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,8192,0.013125333521101209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,10240,0.3528737756941054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,8192,0.018555555078718398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,8192,0.011052444577217102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,7168,0.012135999898115793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,7168,0.017645332548353408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,7168,0.33513156572977704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,7168,0.01036800030204985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,6144,0.01108000013563368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,8192,0.32985422346327037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,6144,0.33376799689398867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,6144,0.009718221922715506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,10240,0.012530666258600024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,5120,0.009711110757456886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,5120,0.015572445260153877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,5120,0.33785332573784727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,5120,0.009111999637550777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,4096,0.009004444711738164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,6144,0.01651733285850949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,4096,0.01483022173245748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,4096,0.008328888979223039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,3584,0.006933333145247565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,4096,0.3175742096371121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,3584,0.014449778530332776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,3072,0.006559999866618051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,3584,0.00793599999613232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,3584,0.3437795639038086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,3072,0.013863110707865821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,2560,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,3072,0.007678222325113084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,3072,0.32528887854682076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,2560,0.013751111096805997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,2048,0.006218666831652324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,2560,0.007323555648326874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,2560,0.33655465973748105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,2048,0.013076444466908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,1536,0.005281777845488654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,2048,0.006999111009968652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,2048,0.33032888836330837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,1536,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,1024,0.00452888881166776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,1536,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,1536,0.3284417788187663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,768,0.00424888895617591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,1024,0.006295111030340195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,1024,0.312187565697564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,768,0.012134222520722283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,512,0.0038897775941424896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,768,0.005981333139869902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,768,0.3396115567949083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,512,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,256,0.0032559999575217566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,512,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,512,0.3246248828040229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,256,0.011694221860832639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,128,0.003248888792263137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,256,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,256,0.33009065522087944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,128,0.011698666546079846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,64,0.00295733329322603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,1024,0.012318222059143914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,1024,128,0.3023200035095215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,1024,32,0.0033973333322339584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,65536,0.02425244450569153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,64,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,1024,32,0.011386666860845355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,65536,0.08850222163730198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,65536,0.30599289470248753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,16384,0.015861334072219003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,16384,0.02858577834235297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,16384,0.3943511115180121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,1024,128,0.005610666755172942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,12288,0.013342222405804528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,16384,0.016522667474216886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,12288,0.022671999202834234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,65536,0.05786489115820991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,12288,0.013594666288958656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,12288,0.36582220925225156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,10240,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,10240,0.012472888661755456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,10240,0.020620443754725985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,8192,0.010332444475756751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,10240,0.347400877210829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,8192,0.018590221802393597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,7168,0.009591110878520543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,8192,0.010818666881985135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,8192,0.31869422064887154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,7168,0.01759733259677887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,7168,0.010426667001512317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,6144,0.0090515555606948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,7168,0.3413848876953125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,6144,0.01657777859105004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,5120,0.008715555899673039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,6144,0.010097777677906884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,6144,0.34773511356777614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,5120,0.015830222103330824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,4096,0.00797333319981893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,5120,0.009095999929640029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,5120,0.3325200080871582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,4096,0.014745778507656522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,3584,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,4096,0.00832088871134652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,4096,0.3382417890760634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,3584,0.014487110906177096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,3072,0.006586666736337874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,3584,0.00794222205877304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,3584,0.3303368886311849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,3072,0.013923555612564087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,2560,0.005642666584915585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,3072,0.007698666718271043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,3072,0.33337065908643937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,2560,0.013746666411558786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,2048,0.005220444252093633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,2560,0.007236444287829929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,2560,0.3288835684458415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,2048,0.01312800000111262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,1536,0.004583999928500918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,2048,0.0069253332912921906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,2048,0.33696889877319336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,1536,0.012796444197495779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,1024,0.0038817777401871155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,1536,0.006649777707126405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,1536,0.32839732699924046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,1024,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,768,0.0036115555299652945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,1024,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,1024,0.33792710304260254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,768,0.012265777422322167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,512,0.0035457776652442086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,768,0.005981333139869902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,768,0.33048799302842885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,512,0.012095111111799875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,256,0.0031679999083280563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,512,0.006019555446174409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,512,0.3363617791069879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,256,0.011713777979214987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,128,0.0028408887899584244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,256,0.005600888695981767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,128,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,256,0.32237333721584743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,64,0.0028728888266616394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,768,128,0.005578666511509154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,768,32,0.0028648889727062653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,64,0.011050666371981302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,65536,0.021730666359265644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,768,32,0.011438222395049201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,65536,0.0868613322575887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,16384,0.012841777669058906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,65536,0.3153759903377957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,65536,0.05577955643335978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,16384,0.027169777287377253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,12288,0.010592888626787396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,768,128,0.30479200681050617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,12288,0.02236622240808275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,12288,0.013821333646774292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,12288,0.3077048990461561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,10240,0.009764444496896531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,10240,0.020459555917316012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,10240,0.012420444852775998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,16384,0.016169778174824186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,10240,0.3908249007331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,16384,0.3096364339192708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,8192,0.018209778600268893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,8192,0.011055111057228513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,7168,0.008316444853941599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,8192,0.34877244631449383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,7168,0.01758044461409251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,6144,0.007610666255156199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,7168,0.010401777923107147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,7168,0.32250844107733834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,6144,0.016484444340070088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,6144,0.009673777553770278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,6144,0.3548666636149089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,5120,0.00795288880666097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,8192,0.009328888522254096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,5120,0.015513777732849121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,4096,0.007344889144102733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,5120,0.31680088573031956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,4096,0.01461155547036065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,4096,0.3162826697031657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,4096,0.008300444318188561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,3584,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,3584,0.014283556077215405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,3584,0.3399493429395888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,3072,0.006654222392373615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,3584,0.0079644446571668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,5120,0.009047110875447592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,3072,0.013825777504179211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,2560,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,3072,0.31660355461968315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,3072,0.007596444752481248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,2560,0.013615999784734515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,2048,0.005208000126812193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,2560,0.007271111011505127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,2560,0.3149671024746365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,2048,0.013091555900043912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,1536,0.004579555657174853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,2048,0.00700800038046307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,2048,0.32272622320387095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,1536,0.012772444221708508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,1024,0.00387999994887246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,1536,0.006650666809744305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,1536,0.32750577396816677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,1024,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,768,0.003573333223660787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,1024,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,1024,0.3253271049923367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,768,0.012177777787049612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,512,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,768,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,768,0.324688884947035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,512,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,256,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,512,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,512,0.32438845104641384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,256,0.011692444483439127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,128,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,256,0.005621333503060871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,256,0.3332800070444743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,128,0.011775111158688864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,64,0.0028568889117903183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,512,128,0.005727111051479976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,512,128,0.30545245276557076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,512,32,0.002855111120475663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,64,0.011707555916574268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,65536,0.019221333993805777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,512,32,0.01110222190618515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,65536,0.08488533231947158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,16384,0.010009778042634329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,65536,0.05008177624808418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,65536,0.33632622824774844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,16384,0.027119111683633592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,16384,0.016456888781653512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,12288,0.008657777474986183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,16384,0.3299493259853787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,12288,0.02235377828280131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,10240,0.009039999710188972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,12288,0.013631111217869652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,12288,0.32004888852437335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,10240,0.02030755579471588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,10240,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,8192,0.00832799987660514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,10240,0.3290719985961914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,8192,0.018556444181336295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,7168,0.00795555528667238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,8192,0.011016888750924004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,8192,0.33446044392055935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,7168,0.017536888519922893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,6144,0.007648000286685095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,7168,0.010378667049937779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,7168,0.3302106592390272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,6144,0.016511110795868766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,5120,0.007963555554548899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,6144,0.009709333380063375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,6144,0.3345386717054579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,5120,0.015558222929636637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,4096,0.007350222104125553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,5120,0.00905777762333552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,5120,0.3303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,3584,0.0070017774899800616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,4096,0.014797333214018079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,4096,0.008295999632941352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,3584,0.014121777481502958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,3072,0.006620444357395172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,4096,0.4742915365431044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,3584,0.32915110058254665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,3072,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,2560,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,3072,0.0076657773719893555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,2560,0.013541333377361298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,3072,0.4283333354526096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,2560,0.3259119987487793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,2048,0.006326222171386083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,2560,0.007234666910436418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,2048,0.013186666700575085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,3584,0.00795733349190818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,1536,0.004350222233268949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,2048,0.0069368887278768755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,2048,0.3797920015123155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,1536,0.0127688886390792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,1536,0.006337777607970768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,1024,0.0038977778620190094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,1536,0.33431198861863876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,1024,0.012176888684431711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,768,0.003670222229427762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,1024,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,1024,0.3298284477657742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,768,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,768,0.006242666807439592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,512,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,768,0.3263813389672173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,512,0.01180355581972334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,256,0.003176888864901331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,512,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,512,0.36496443218655056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,256,0.011700444751315646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,128,0.002903999967707528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,256,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,256,0.3371715545654297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,128,0.011697777443461947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,64,0.0028311111446883944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,256,128,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,256,32,0.0026835555003749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,256,128,0.3247777885860867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,64,0.01146399974822998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,256,32,0.011347555451922946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,65536,0.08207733101314969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,65536,0.3422844409942627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,65536,0.048525333404541016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,16384,0.009324444664849175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,16384,0.026808889375792608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,16384,0.3236408763461643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,16384,0.016576889488432143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,12288,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,65536,0.014733332726690503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,12288,0.02228533393806881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,10240,0.007707555260923173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,12288,0.013425778183672162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,12288,0.3236622280544705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,10240,0.020543111695183646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,10240,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,8192,0.0069422221018208405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,10240,0.3385048972235785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,8192,0.01861422260602315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,7168,0.006604444649484422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,8192,0.010728889041476779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,8192,0.33669066429138184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,7168,0.017543110582563613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,7168,0.0103502223889033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,7168,0.33060354656643337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,6144,0.01629333363638984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,6144,0.009761778016885122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,6144,0.3250124454498291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,5120,0.006311111152172089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,5120,0.015015999476114908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,5120,0.33418843481275773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,4096,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,6144,0.006684444430801604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,4096,0.014783999986118741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,4096,0.00833155545923445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,3584,0.006249777972698212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,4096,0.30272089110480416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,3584,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,3584,0.33114398850335014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,3584,0.008000888758235509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,3072,0.006246222390068903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,5120,0.008990222381220924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,3072,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,2560,0.006309333360857434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,3072,0.007711110843552484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,3072,0.3145413398742676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,2048,0.006251555350091722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,2560,0.0135057775510682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,2560,0.007339555356237624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,2560,0.33137332068549263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,2048,0.01315199997689989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,1536,0.004560889055331548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,2048,0.3360622194078233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,1536,0.012806221842765808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,1024,0.0038693332009845306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,1536,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,1536,0.3253840075598823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,1024,0.012446222205956778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,768,0.0038515557017591265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,1024,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,1024,0.3224906656477186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,768,0.012096000214417776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,512,0.00350222239891688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,768,0.006223111102978389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,768,0.29823467466566295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,512,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,512,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,512,0.3322826756371392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,2048,0.0070462218589252895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,256,0.002833777831660377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,256,0.011667555405033959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,256,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,128,0.0029528888149393927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,256,0.3207155598534478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,128,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,64,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,96,128,128,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,128,32,0.002850666642189026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,96,128,128,0.31254755126105416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,65536,0.01313866674900055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,64,0.011448000040319232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,16384,0.00796088907453749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,128,32,0.01146933353609509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,12288,0.007663999994595845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,65536,0.08146755562888251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,10240,0.007320889168315464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,12288,0.02239377796649933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,8192,0.0070088886552386815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,10240,0.020265777905782063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,7168,0.006956444846259222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,8192,0.018363555272420246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,6144,0.006671111202902264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,7168,0.017232000827789307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,5120,0.0063271112740039825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,6144,0.016122667325867545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,4096,0.005909333212508096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,5120,0.015398222539159985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,3584,0.006346666564544042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,4096,0.014461333552996317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,3072,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,16384,0.02683111031850179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,2560,0.00656444455186526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,3072,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,2048,0.005971555494599872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,2560,0.013501333693663279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,1536,0.004528000122971005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,2048,0.013155555559529198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,1024,0.0038737778862317405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,1536,0.012823111481136747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,768,0.0035457776652442086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,1024,0.012120000190205045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,512,0.0034826666944556763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,768,0.012409778104888068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,256,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,512,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,128,0.002938666691382726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,3584,0.014180445008807711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,256,0.011651555697123209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,64,0.0025840000145965153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,64,32,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,128,0.01145244472556644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,64,0.011496889094511667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,65536,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,16384,0.00701422244310379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,65536,0.08100533485412598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,12288,0.007623111208279927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,16384,0.026750221848487854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,10240,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,12288,0.022287110487620037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,8192,0.006937777830494775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,10240,0.020277332928445604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,7168,0.0069226668112807805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,8192,0.018588445252842374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,6144,0.006568000134494569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,7168,0.01720355616675483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,5120,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,6144,0.016509332590632968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,4096,0.00590133335855272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,64,32,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,3584,0.006326222171386083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,4096,0.014841778410805596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,3072,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,3584,0.014205333259370593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,2560,0.006291555447710886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,3072,0.013863110707865821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,2048,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,2560,0.013456000222100152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,1536,0.004496888981925117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,2048,0.013085333009560903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,1024,0.003855111284388436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,1536,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,768,0.003572444534964032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,1024,0.012392889294359418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,512,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,5120,0.015268445014953613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,256,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,512,0.012055111428101858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,128,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,256,0.011448889142937131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,64,0.0028391112056043413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,128,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,96,32,32,0.0027502222607533136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,64,0.011369778050316704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,32,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,16384,0.34916265805562335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,16384,0.2113893296983507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,96,32,768,0.012120889292822944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,12288,0.2602328989240858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,16384,0.13870933320787218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,12288,0.1645760006374783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,12288,0.11262667179107666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,16384,0.3490560054779053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,10240,0.20715022087097168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,10240,0.13723822434743246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,12288,0.30666311581929523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,8192,0.16955999533335367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,8192,0.11715822749667698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,10240,0.3496862252553304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,8192,0.07427466577953763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,8192,0.3409599992964003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,7168,0.09899555312262641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,10240,0.09383289019266765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,7168,0.14693955580393472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,7168,0.06431377596325345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,7168,0.33616267310248477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,6144,0.08763111299938625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,6144,0.06528711318969727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,5120,0.10603288809458415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,6144,0.33444976806640625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,6144,0.1266382270389133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,5120,0.07509599791632758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,4096,0.08823288811577691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,5120,0.3332560062408447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,5120,0.04751822352409363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,4096,0.06890844636493258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,3584,0.0772648917304145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,3584,0.05638577540715536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,4096,0.039416889349619545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,3072,0.06579377916124132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,4096,0.32194044854905873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,3584,0.041654222541385226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,3584,0.3228488763173421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,2560,0.056532442569732666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,3072,0.05013688736491733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,3072,0.032756444480684065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,2560,0.04520533482233683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,3072,0.32029422124226886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,2048,0.047154668304655284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,2048,0.03962755534383986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,2560,0.4048995441860623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,1536,0.0358586675590939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,2048,0.30407023429870605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,2048,0.024854222933451336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,1536,0.04324000080426534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,1536,0.020976000361972388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,1024,0.02583377725548214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,1536,0.31219911575317383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,1024,0.031722668144438006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,768,0.02028888960679372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,1024,0.015083556373914083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,2560,0.02835644284884135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,1024,0.35888534122043186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,512,0.013971555564138623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,768,0.03161244591077169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,768,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,768,0.32121067576938206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,512,0.023080888721677992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,256,0.009379555781682333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,512,0.011678222152921887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,512,0.337768899069892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,256,0.010421333213647207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,256,0.018410666121376883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,256,0.3186844454871284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,128,0.017277333471510146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,64,0.006579555571079254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,65536,128,0.010678222609890832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,65536,128,0.27339288923475474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,32,0.007074666519959767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,64,0.01843377782238854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,65536,32,0.02171733313136631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,65536,0.30175556076897514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,65536,0.2490231196085612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,16384,0.07969511217541166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,65536,128,0.007368000017272101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,16384,0.07537155681186251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,65536,0.1633626619974772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,12288,0.0620959997177124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,16384,0.045329779386520386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,12288,0.05807999769846598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,16384,0.3047991063859728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,10240,0.053675555520587497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,12288,0.03734933336575826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,10240,0.04936266607708401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,12288,0.3135555585225423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,65536,0.32440445158216685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,8192,0.043041778935326465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,10240,0.03134222163094415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,10240,0.3261191050211589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,7168,0.038592888249291316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,8192,0.027600000301996868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,8192,0.31314312087164986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,7168,0.03836799992455377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,6144,0.03492177857293023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,7168,0.022867555419603985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,7168,0.3063982327779134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,6144,0.034541332059436373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,6144,0.33202756775750053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,6144,0.02145333256986406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,5120,0.029021332661310833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,8192,0.04418222109476725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,5120,0.03127911024623447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,4096,0.025070221887694463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,5120,0.2797786659664578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,4096,0.02908266584078471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,4096,0.015605333778593274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,4096,0.29829332563612193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,3584,0.021723555194007024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,3584,0.024526221884621516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,3584,0.3420666588677301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,3072,0.018772444791264005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,3584,0.015402666396564908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,3072,0.02257333364751604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,5120,0.018922666708628338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,2560,0.016537777251667447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,3072,0.33362756835089785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,3072,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,2560,0.020629333125220407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,2560,0.011952888634469775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,2048,0.015421332584487068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,2048,0.019277332557572257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,2048,0.01108444482088089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,1536,0.011110222174061669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,1536,0.01976977785428365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,1536,0.34580267800225156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,2560,0.31811022758483887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,1024,0.008040889269775814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,1536,0.009166222479608323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,1024,0.01715022159947289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,2048,0.3362897766960992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,768,0.0069360000391801195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,1024,0.007987555530336168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,1024,0.3343831168280707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,768,0.01513777838812934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,512,0.005218666460778978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,768,0.007304000357786815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,768,0.3248151143391927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,512,0.014368888404634265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,256,0.003891555385457145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,512,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,512,0.3285484313964844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,256,0.01274577776590983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,128,0.0037946667936113146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,256,0.006680000159475539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,256,0.3185359901852078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,128,0.012416889270146688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,64,0.0035182221068276297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,16384,128,0.006642666541867786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,16384,128,0.3101360003153483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,16384,32,0.003338666632771492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,64,0.013501333693663279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,16384,32,0.014519999424616495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,65536,0.23499377568562826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,65536,0.19867911603715685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,16384,0.06929510831832886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,65536,0.14153156015608045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,16384,0.06085866689682007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,16384,0.04127022292878892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,12288,0.05396977729267544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,16384,0.3277555571662055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,65536,0.34323999616834855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,12288,0.04796444376309713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,10240,0.04551466637187534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,12288,0.3056453333960639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,10240,0.04130399889416165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,8192,0.03793599870469835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,10240,0.027694221999910142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,10240,0.3208551141950819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,8192,0.025588444537586633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,7168,0.03334133492575751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,8192,0.31506845686170787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,7168,0.032313777340783015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,7168,0.3130106661054823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,7168,0.020604444874657523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,12288,0.03253600001335144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,6144,0.02936799989806281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,6144,0.3056737846798367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,6144,0.02071822186311086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,5120,0.025253333979182776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,8192,0.037117332220077515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,5120,0.025470222036043804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,4096,0.022364444202846948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,5120,0.016319110989570618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,5120,0.31385599242316353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,4096,0.023892444041040208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,4096,0.01428533262676663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,6144,0.029763556189007227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,3584,0.02142222225666046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,3584,0.013850666582584381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,3584,0.29734844631618923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,3072,0.016171556380059984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,3072,0.020634666085243225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,3072,0.01219911128282547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,3072,0.3320666684044732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,2560,0.014115555418862237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,2560,0.01927199959754944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,3584,0.018623999423450895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,4096,0.32416799333360463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,2048,0.0131217779384719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,2560,0.3368453449673123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,2560,0.01108444482088089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,2048,0.017679999272028606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,2048,0.010237333675225576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,1536,0.010235555469989777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,2048,0.32393866115146214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,1536,0.018231999542978074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,1024,0.007984889050324758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,1536,0.008984000318580205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,1536,0.3448933230506049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,1024,0.015127110812399121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,768,0.0069466663731469055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,1024,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,1024,0.3260604540506999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,768,0.013996443814701505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,512,0.004892444445027245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,768,0.007288888924651676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,768,0.33704710006713867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,512,0.013752889302041797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,256,0.003924444317817688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,512,0.0068986668354935115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,512,0.32624178462558323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,256,0.012471111284361945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,128,0.003536000019974179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,256,0.006588444527652528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,256,0.3280044396718343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,128,0.011980444192886353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,64,0.0035706667436493766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,12288,128,0.2993866602579753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,12288,128,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,12288,32,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,64,0.0125937776433097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,12288,32,0.013818666338920593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,65536,0.20345600446065268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,65536,0.19226577546861437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,16384,0.05667466587490506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,65536,0.13612711429595947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,16384,0.059028446674346924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,16384,0.040018667777379356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,12288,0.04420622189839681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,16384,0.3344568941328261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,65536,0.3407199912601047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,12288,0.0466213325659434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,12288,0.030560887522167627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,10240,0.03781422310405307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,12288,0.3204675515492757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,10240,0.03997688823276096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,10240,0.026016889346970454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,10240,0.3323591020372179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,8192,0.03620355659061008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,8192,0.022668444448047217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,7168,0.028213332096735638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,7168,0.031812445984946355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,8192,0.03192889028125339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,7168,0.3351324399312337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,6144,0.02493422230084737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,7168,0.01943911115328471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,6144,0.027450667487250432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,8192,0.32676532533433705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,5120,0.020830222302012973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,6144,0.018554665976100497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,6144,0.33185601234436035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,5120,0.024541333317756653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,4096,0.018177777528762817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,5120,0.01590399940808614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,5120,0.32357600000169545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,4096,0.023580445183648005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,3584,0.01564800077014499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,4096,0.013682666752073499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,3584,0.020993777447276648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,3584,0.013562666873137156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,3072,0.013703111145231458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,4096,0.4677422311570909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,3584,0.3166915575663249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,2560,0.012187555432319641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,3072,0.020282665888468426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,3072,0.011807111402352652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,2560,0.019117333822780185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,3072,0.32420534557766384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,2048,0.011818666425016193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,2560,0.010469333165221743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,2560,0.3411448796590169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,2048,0.017198221551047433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,1536,0.009076444639099969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,2048,0.00905688852071762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,2048,0.326819552315606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,1024,0.0070773329999711775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,1536,0.008020444048775567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,1536,0.3313884470197889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,1024,0.015108444624476962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,768,0.006127111199829314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,1024,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,768,0.014801777071423002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,1024,0.35500444306267637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,512,0.004914666629499859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,768,0.0069511110583941145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,768,0.3213386535644531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,512,0.01349955548842748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,512,0.3200017876095242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,512,0.006688888702127669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,256,0.0042133331298828125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,1536,0.016944888565275405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,256,0.012505778008037143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,256,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,128,0.003580444388919406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,256,0.3293164571126302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,128,0.011736888852384357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,64,0.003522666792074839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,10240,128,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,10240,128,0.31947467062208385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,64,0.012426666915416718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,10240,32,0.01348088930050532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,65536,0.15868177678849962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,65536,0.15094667010837132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,65536,0.07849778069390191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,16384,0.04563644528388977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,10240,32,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,65536,0.3363795545366075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,16384,0.04898133211665683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,12288,0.03658577799797058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,16384,0.33410578303866917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,16384,0.024482667446136475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,12288,0.03850488861401876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,10240,0.0310533344745636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,12288,0.018538667096032035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,12288,0.3198853333791097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,10240,0.03319733341534933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,8192,0.025428444147109985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,10240,0.01685688893000285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,10240,0.32766132884555393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,8192,0.029711110724343195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,7168,0.022870222727457683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,8192,0.015294222368134392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,8192,0.31621599197387695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,7168,0.02581866747803158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,6144,0.020696888367335003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,7168,0.013443555268976422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,7168,0.3274746735890706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,6144,0.023445333043734234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,6144,0.01239999963177575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,5120,0.017682666579882305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,6144,0.3173351022932264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,5120,0.021533333592944678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,4096,0.015607110328144498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,5120,0.2947759893205431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,4096,0.020647111866209242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,4096,0.309803565343221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,4096,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,3584,0.01867288847764333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,3584,0.3383333418104384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,5120,0.011722666521867117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,3072,0.011720000041855706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,3584,0.009330666727489894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,3072,0.018279999494552612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,3584,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,3072,0.00849333322710461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,2560,0.010798222488827176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,2560,0.01724177764521705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,2560,0.3285111056433784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,2560,0.008034666379292807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,2048,0.00981422182586458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,2048,0.016179554992251925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,2048,0.32829777399698895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,1536,0.00776800016562144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,2048,0.007633777956167857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,1536,0.01595288846227858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,1536,0.007179555793603261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,1024,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,3072,0.32365510198805064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,1024,0.014459555347760519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,1024,0.006384000182151794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,768,0.004895110925038655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,1536,0.4752515686882867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,1024,0.33217954635620117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,512,0.0041893331540955436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,768,0.01313066648112403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,768,0.006321777900060018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,768,0.32327644030253094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,512,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,512,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,512,0.3224426640404595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,256,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,128,0.003567999849716822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,256,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,128,0.011735999749766456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,256,0.38997332255045575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,64,0.003028444531891081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,8192,128,0.0058746664888328975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,32,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,64,0.012128888732857175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,8192,32,0.01314399970902337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,65536,0.14791556199391684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,8192,256,0.0035502223504914176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,65536,0.14625867207845053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,16384,0.045775112178590566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,65536,0.07601600223117404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,16384,0.04715377754635281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,8192,128,0.2977866596645779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,16384,0.3295973406897651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,65536,0.37596978081597227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,12288,0.033681776788499616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,16384,0.02344977855682373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,12288,0.03754311137729221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,12288,0.0178195552693473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,10240,0.02888266576661004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,12288,0.31842756271362305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,10240,0.0317555566628774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,8192,0.023349333140585158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,10240,0.37188087569342715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,8192,0.028233776489893597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,8192,0.01458666721979777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,8192,0.35016711552937824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,7168,0.021372444099850122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,7168,0.02534399926662445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,7168,0.3355662292904324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,6144,0.019344889455371432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,7168,0.01314133322901196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,10240,0.0162782222032547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,6144,0.022755554980701868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,6144,0.012172444827026792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,5120,0.016007110476493835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,6144,0.3585057788425022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,5120,0.02130311065249973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,4096,0.014519999424616495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,5120,0.011284444067213269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,5120,0.3302826616499159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,4096,0.020599999361568026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,4096,0.00943199959066179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,4096,0.34265510241190594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,3584,0.018607111440764535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,3584,0.008620444271299574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,3584,0.3324524561564128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,3072,0.011261333194043903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,3072,0.018079999420377944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,3072,0.34329689873589414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,3072,0.007994666695594788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,2560,0.010384889112578498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,3584,0.012834666503800286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,2560,0.016858667135238647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,2048,0.009722666607962714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,2560,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,2560,0.33713245391845703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,2048,0.01518311103185018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,1536,0.007681777907742395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,2048,0.007048889166778988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,2048,0.32766132884555393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,1536,0.015855110353893705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,1024,0.005636444522274866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,1536,0.006888000087605582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,1536,0.3279208872053358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,1024,0.014184888866212634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,768,0.004577777865860197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,1024,0.00630666646692488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,1024,0.329366233613756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,768,0.01387733303838306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,512,0.0041759999261962045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,768,0.005974222388532426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,768,0.3333760102589925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,512,0.012863111164834766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,256,0.003565333369705412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,512,0.005912000106440649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,512,0.3348284562428792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,256,0.012396444049146442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,128,0.003257777748836411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,256,0.005712000032265981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,256,0.31847111384073895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,128,0.011715555356608497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,7168,128,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,64,0.0032053333189752367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,7168,32,0.0035173334181308746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,7168,128,0.27768445014953613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,64,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,7168,32,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,65536,0.1430435578028361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,16384,0.03817155626085069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,65536,0.07308266560236613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,16384,0.045816888411839805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,16384,0.022478222846984863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,16384,0.32581154505411786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,65536,0.33207644356621635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,12288,0.03165333469708761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,12288,0.03683022326893277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,12288,0.01737866633468204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,10240,0.02749866743882497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,65536,0.12014488379160564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,12288,0.3125671015845405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,8192,0.02202133337656657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,10240,0.031594667169782854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,10240,0.015791111522250705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,10240,0.3216613398657905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,8192,0.02721066607369317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,7168,0.018833777970737882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,8192,0.013884444203641681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,8192,0.3359368907080756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,7168,0.024359110328886244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,7168,0.01276000009642707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,7168,0.3347511026594374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,6144,0.022509333160188463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,6144,0.333621342976888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,6144,0.011799111134476133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,5120,0.015041778484980265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,5120,0.020710221595234342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,5120,0.3183697859446208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,4096,0.013629333012633853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,5120,0.010772444307804108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,6144,0.018000889155599806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,4096,0.3410693274603949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,4096,0.009013333254390294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,3584,0.011374221907721626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,3584,0.018612444400787354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,3584,0.32861243353949654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,3584,0.00830844458606508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,3072,0.010289777484205034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,3072,0.017656889226701524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,3072,0.3319724400838216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,3072,0.007956444389290279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,2560,0.009468444519572789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,2560,0.016234666109085083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,2560,0.3231377866533068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,2560,0.007666666474607255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,2048,0.01483022173245748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,2048,0.31909600893656415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,4096,0.020062221421135795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,1536,0.007092444433106317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,2048,0.007272889216740926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,1536,0.014864888456132678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,2048,0.009109333157539368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,1536,0.32640443907843697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,1024,0.005558222118351195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,1536,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,1024,0.014120888378885059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,1024,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,768,0.004896000027656555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,1024,0.3031795554690891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,768,0.013119999733236102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,512,0.0042462220622433555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,768,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,768,0.3265084425608317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,512,0.012856889102194043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,256,0.003544888976547453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,512,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,512,0.3218604458702935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,256,0.012094222009181976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,256,0.005894222193294101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,256,0.3304719924926758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,128,0.011404444773991903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,6144,128,0.3050417900085449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,6144,128,0.005766222046481238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,64,0.0028924445311228433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,32,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,64,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,6144,32,0.01276088919904497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,65536,0.10457688570022583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,65536,0.14033599694569907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,6144,128,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,16384,0.0335804455810123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,65536,0.07090133428573608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,16384,0.0443742209010654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,65536,0.3435377809736464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,16384,0.020980444219377305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,12288,0.02778844369782342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,16384,0.3315644529130724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,12288,0.03408622079425388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,10240,0.024864888853496973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,12288,0.27946755621168345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,12288,0.01696622206105126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,8192,0.022163555026054382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,10240,0.02937155630853441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,10240,0.015398222539159985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,10240,0.2974710994296604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,7168,0.019342222147517733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,8192,0.013723555538389417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,8192,0.026708443959554035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,8192,0.3308124542236328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,6144,0.017872000734011333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,7168,0.023933332827356126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,7168,0.012455110748608908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,7168,0.3194631205664741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,6144,0.022456000248591106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,6144,0.011485333244005838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,6144,0.3166168795691596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,5120,0.02075822154680888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,4096,0.011490667031870948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,5120,0.010024000373151567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,4096,0.019859555694792006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,5120,0.38785955641004777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,3584,0.010206222534179688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,4096,0.3189048767089844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,3584,0.018036444981892902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,3584,0.31349065568712026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,3584,0.008018666671382057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,3072,0.010335110955768162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,5120,0.012869333227475485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,3072,0.007689778175618913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,2560,0.009012444151772393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,3072,0.3054951032002767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,2560,0.015923556354310777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,2560,0.3069244490729438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,4096,0.00834222220712238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,2048,0.008951110972298516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,2560,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,2048,0.014564444621404013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,2048,0.007259555160999298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,2048,0.3319297896491157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,1536,0.007260444263617198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,3072,0.01696888936890496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,1536,0.014473777678277759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,1024,0.00526577772365676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,1536,0.33184978697035045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,1024,0.013438222308953604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,1024,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,1024,0.31771111488342285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,768,0.004568888909286923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,768,0.013094222380055321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,768,0.34534488783942324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,768,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,512,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,1536,0.006654222392373615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,512,0.012735111018021902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,256,0.0038782221575578055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,512,0.3239475621117486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,512,0.00590133335855272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,256,0.012144000165992312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,128,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,256,0.00562755556570159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,256,0.32005778948465985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,128,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,64,0.0031537777847713898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,5120,128,0.005895111295912001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,5120,128,0.3053493234846327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,5120,32,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,64,0.012449777788586087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,65536,0.08396533462736343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,5120,32,0.012744888663291931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,65536,0.0999644464916653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,16384,0.02786933382352193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,65536,0.06832977798249987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,16384,0.032940444019105695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,65536,0.33173155784606934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,16384,0.3248666657341851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,12288,0.024819556209776137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,12288,0.02646488944689433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,12288,0.32706578572591144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,10240,0.021298666795094807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,10240,0.02299022177855174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,10240,0.3309946589999729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,16384,0.02030933399995168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,8192,0.015858666764365304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,10240,0.014907555447684394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,8192,0.021645334031846788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,12288,0.016584888100624084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,7168,0.014123555686738757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,8192,0.3326115608215332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,7168,0.019647111495335896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,7168,0.321706665886773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,7168,0.01165511127975252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,6144,0.018598222070270114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,6144,0.3402408758799235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,8192,0.01311644415060679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,5120,0.011286222272449069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,6144,0.010695111420419482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,5120,0.017401778035693698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,6144,0.013519110778967539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,4096,0.010581333604123859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,5120,0.31896177927652997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,4096,0.016011555989583332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,4096,0.008329778081840938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,3584,0.009014222357008193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,4096,0.2957875463697645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,3584,0.0155404441886478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,3072,0.00868888861603207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,3584,0.007975999679830339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,3584,0.3272622161441379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,3072,0.014560888210932413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,2560,0.007968000239796108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,3072,0.007692444655630324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,3072,0.3089146614074707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,2560,0.013793778088357715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,2048,0.006622222148709827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,2560,0.007296889192528195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,5120,0.009692444569534725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,2560,0.32728976673550075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,1536,0.0048942222363419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,2048,0.013487111363146039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,2048,0.007038222418891058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,1536,0.013447999954223633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,2048,0.3086453278859456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,1024,0.004190222256713443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,1536,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,1536,0.29283377859327525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,1024,0.013070222404268054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,768,0.004223111189073986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,1024,0.006326222171386083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,1024,0.3131626711951362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,768,0.012807110945383707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,512,0.0038248888320393036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,768,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,768,0.31980978118048775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,512,0.012464000119103326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,256,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,512,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,512,0.3337048954433865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,256,0.011782222323947482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,256,0.005673777725961473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,128,0.0032222221295038858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,256,0.3293386565314399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,128,0.011387555963463254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,64,0.002862222285734283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,4096,128,0.005584888988071018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,4096,128,0.31310489442613393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,4096,32,0.003176888864901331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,64,0.011390222443474663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,65536,0.07331822315851848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,4096,32,0.012286221815480126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,65536,0.09445866611268784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,16384,0.02490933405028449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,65536,0.34626489215426975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,16384,0.30849332279629177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,16384,0.01998666591114468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,12288,0.022118222382333543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,65536,0.06759111086527507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,12288,0.02570399973127577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,16384,0.031155556440353394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,12288,0.016452444924248587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,10240,0.020271110865804885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,12288,0.30558133125305176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,10240,0.022663111488024395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,10240,0.014730667074521383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,8192,0.017982222967677645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,10240,0.33276089032491046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,8192,0.021073778470357258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,7168,0.016224000189039443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,8192,0.012608889076444836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,8192,0.33914934264289004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,7168,0.019398222366968792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,6144,0.014349333114094205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,7168,0.011325333681371478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,7168,0.3264942169189453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,6144,0.018233777748213876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,6144,0.010767999622556897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,6144,0.3465786774953206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,5120,0.0169840008020401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,5120,0.009323555562231276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,4096,0.009940444595283931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,5120,0.32780443297492134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,4096,0.015132443772421943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,4096,0.33726223309834796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,4096,0.008352888955010308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,3584,0.008960888617568547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,5120,0.010839111275143094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,3584,0.014450665977266101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,3072,0.008015111088752747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,3584,0.33211734559800893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,3072,0.014134221606784396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,3072,0.007640889121426477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,2560,0.007328888608349695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,3072,0.34366488456726074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,2560,0.01350488927629259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,2560,0.33381422360738117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,2560,0.007328000333574083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,2048,0.01311466677321328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,3584,0.007961778177155389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,1536,0.004892444445027245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,2048,0.007096889118353526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,2048,0.47651465733846027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,1536,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,1024,0.004281777888536453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,1536,0.006623111251327727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,1536,0.32899289660983616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,1024,0.01276088919904497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,768,0.003922666526503033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,1024,0.00627377794848548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,1024,0.31369243727789986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,2048,0.0069066666894488865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,768,0.012796444197495779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,512,0.00360355567600992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,768,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,512,0.01240355521440506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,256,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,512,0.005967111223273807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,512,0.3075084421369764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,256,0.011753777662913004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,128,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,256,0.005626666463083691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,256,0.32811466852823895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,128,0.011391999820868174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,64,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3584,128,0.005763555566469829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3584,32,0.0031431110368834603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,64,0.011773332953453064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,65536,0.06591733296712239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3584,32,0.012151999606026543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,768,0.31619466675652397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,65536,0.09563999705844456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,16384,0.022544888986481562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,65536,0.06577066580454509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,16384,0.03071111109521654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,65536,0.33848267131381565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,16384,0.30871378050910103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,12288,0.019609777463806998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,16384,0.01976977785428365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3584,128,0.30066222614712185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,12288,0.024400000770886738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,12288,0.016231111354298063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,10240,0.016581333345837064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,12288,0.33676711718241376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,10240,0.022447110878096685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,8192,0.014813333749771118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,10240,0.01441333360142178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,10240,0.322559118270874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,8192,0.020433778564135235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,7168,0.012912888493802814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,8192,0.012125333150227865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,8192,0.3322800000508626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,7168,0.019193778435389202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,7168,0.010406222608354356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,6144,0.011931555966536203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,7168,0.3170657687717014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,6144,0.018049778209792245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,6144,0.00983199973901113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,5120,0.010454222559928894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,6144,0.3424622217814128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,5120,0.01681866745154063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,4096,0.009425777528021071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,5120,0.009367110828558603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,5120,0.33414400948418516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,4096,0.0148026661740409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,3584,0.008346666892369589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,4096,0.008346666892369589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,4096,0.3317564328511556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,3584,0.01421333352724711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,3072,0.0079644446571668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,3584,0.008025777836640676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,3584,0.3430195649464925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,3072,0.014187556174066333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,2560,0.00700800038046307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,3072,0.007676444119877285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,3072,0.3372551070319281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,2560,0.013419555293189155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,2048,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,2560,0.007348444726732042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,2560,0.34061866336398655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,2048,0.013109332985348172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,1536,0.004907555464241239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,2048,0.0070062221752272705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,2048,0.32715021239386666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,1536,0.013080889152155982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,1024,0.004266666869322459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,1536,0.006659555352396435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,1536,0.2952951060401069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,1024,0.012433778080675336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,768,0.00416088890698221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,1024,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,1024,0.3230835596720378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,768,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,512,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,768,0.005945777727497949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,768,0.3203528987036811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,512,0.012116444607575735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,256,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,512,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,512,0.3257804446750217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,256,0.01163822246922387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,128,0.0032115555885765287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,256,0.00564444437623024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,256,0.3224266635047065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,128,0.0110880004035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,64,0.0031893334041039148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,3072,128,0.005904888941182031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,3072,32,0.0028880000528362063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,3072,128,0.30813243654039174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,65536,0.055758224593268506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,64,0.011036444041464063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,3072,32,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,65536,0.09152088562647502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,16384,0.020648888415760465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,65536,0.06476266516579522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,65536,0.2954275608062744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,16384,0.029656890365812514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,12288,0.018593778212865193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,16384,0.019650666250122916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,16384,0.3390728897518582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,12288,0.024336889386177063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,12288,0.015311110350820752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,10240,0.015461333923869662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,12288,0.31067734294467503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,10240,0.021996445126003687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,8192,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,10240,0.3491475582122803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,8192,0.02015911042690277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,7168,0.012831110921170978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,8192,0.011159111228254108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,8192,0.3266399966345893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,7168,0.018960000740157235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,7168,0.01036977767944336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,6144,0.013039111263222165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,7168,0.3387208779652913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,6144,0.017415111263593037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,10240,0.013748444616794586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,5120,0.010000000397364298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,6144,0.4474773406982422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,5120,0.01550044450494978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,5120,0.009312888814343346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,5120,0.4058115482330322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,4096,0.009017777939637503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,4096,0.015417777829700045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,4096,0.3374195628696018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,3584,0.008246222303973304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,4096,0.008307555483447181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,6144,0.009698666632175446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,3584,0.014536889062987434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,3072,0.0076453329788313965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,3584,0.008006222546100616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,3584,0.3593404557969835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,3072,0.014030221435758801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,2560,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,3072,0.007658667034573025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,3072,0.32808176676432294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,2560,0.013472889032628802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,2048,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,2560,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,2560,0.35458845562405056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,2048,0.013022222452693515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,1536,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,2048,0.3377564483218723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,1536,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,1536,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,1024,0.004218666503826777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,1024,0.012455110748608908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,1024,0.329457786348131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,2048,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,1024,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,768,0.0038906666967603895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,768,0.012127111355463663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,1536,0.3309564325544569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,768,0.3090248902638753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,512,0.003567999849716822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,768,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,512,0.012120889292822944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,512,0.0058124446206622655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,256,0.0035146665241983202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,512,0.33316267861260307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,256,0.011739555332395764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,128,0.003177777760558658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,256,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,256,0.33393510182698566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,128,0.011043555206722684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,64,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2560,128,0.005568888866239124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2560,32,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2560,128,0.31229599316914874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,65536,0.045884443653954395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,64,0.011130666567219628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2560,32,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,65536,0.0873111089070638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,16384,0.018636445204416912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,65536,0.34037333064609104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,16384,0.02921688887808058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,16384,0.018908444378111098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,16384,0.3163164456685384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,12288,0.014367111855083041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,12288,0.02362222141689724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,10240,0.013844444519943662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,12288,0.3326542112562391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,12288,0.015227556228637695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,10240,0.021381333470344543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,65536,0.06367733081181844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,10240,0.012427555190192329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,10240,0.3278462092081706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,8192,0.019621334142155118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,7168,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,8192,0.01107022249036365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,8192,0.40352267689175075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,7168,0.33844712045457626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,7168,0.010375111467308467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,6144,0.010684444672531553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,6144,0.016926222377353244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,8192,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,6144,0.31946754455566406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,5120,0.010067555639478896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,6144,0.009752000371615091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,5120,0.015409777561823526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,4096,0.009992889232105678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,5120,0.3322862254248725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,4096,0.014880888991885714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,4096,0.00832711077398724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,3584,0.009150221943855286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,3584,0.01418133411142561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,3584,0.3484773370954726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,5120,0.009259555902745988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,3072,0.00830577810605367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,3584,0.007975999679830339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,7168,0.017871111631393433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,3072,0.014083556003040738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,2560,0.007332444190979004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,3072,0.007679111427730984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,4096,0.32304978370666504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,3072,0.33361421691046816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,2560,0.013495999905798169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,2048,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,2560,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,2560,0.31616711616516113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,2048,0.01313777764638265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,1536,0.004943999979231092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,2048,0.0069875551594628235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,2048,0.3330399990081787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,1536,0.01277511070171992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,1024,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,1536,0.006684444430801604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,1536,0.31500177913241917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,1024,0.012510221865442065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,768,0.003913777983850903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,1024,0.33518044153849286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,768,0.012348444097571902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,768,0.005936000082227919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,768,0.32372623019748265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,512,0.012068444656001197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,1024,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,512,0.34656622674730087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,256,0.003532444437344869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,512,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,256,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,512,0.003591999825504091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,256,0.005678222411208683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,128,0.0029351111087534162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,256,0.33579643567403156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,128,0.011385777758227454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,2048,128,0.005585777676767773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,64,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,2048,32,0.0029013332807355454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,2048,128,0.3245697816212972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,65536,0.03868088788456387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,32,0.01204533295498954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,65536,0.08827377690209283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,65536,0.34150134192572695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,16384,0.014599111345079211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,16384,0.02885422110557556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,2048,64,0.011789333489206104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,16384,0.35097866588168675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,16384,0.018607111440764535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,12288,0.02363377809524536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,65536,0.06200177801979912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,12288,0.013586666848924426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,12288,0.33146222432454425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,10240,0.011704000333944956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,10240,0.021613332960340712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,10240,0.3539537853664822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,10240,0.01258044441541036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,8192,0.012097777591811286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,12288,0.013569778038395775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,8192,0.018583999739752877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,7168,0.010958222051461538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,8192,0.33945155143737793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,7168,0.017253332667880587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,7168,0.01036444471942054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,6144,0.01109155515829722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,7168,0.32177334361606175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,6144,0.016339555382728577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,6144,0.30601954460144043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,5120,0.009536888864305284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,6144,0.009695111049546136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,8192,0.011008000208271874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,5120,0.015466666883892484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,4096,0.008278222547637092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,5120,0.32748889923095703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,5120,0.009029332962301042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,3584,0.00737955586777793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,4096,0.014821334017647637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,4096,0.008303110798199972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,3584,0.014189332723617554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,3072,0.006973333656787872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,3584,0.007939555578761632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,4096,0.4556293487548828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,3584,0.32724889119466144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,2560,0.005993777679072485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,3072,0.013809777796268463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,3072,0.007631111476156447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,3072,0.3228515519036187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,2560,0.01351377781894472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,2048,0.00526577772365676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,2560,0.007101332975758447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,2048,0.01311822235584259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,1536,0.004914666629499859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,2048,0.0070017774899800616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,2048,0.321399105919732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,1536,0.01277599980433782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,1024,0.004190222256713443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,1536,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,1024,0.0123848890264829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,1536,0.36510488722059464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,768,0.0038773334688610504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,1024,0.3125760025448269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,1024,0.006217777729034424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,768,0.012075555821259817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,512,0.003535110917356279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,768,0.00600355573826366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,768,0.30117689238654244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,2560,0.33722400665283203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,512,0.01180355581972334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,256,0.0031822222388452957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,512,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,512,0.3495982223086887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,256,0.011601777540312873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,128,0.0031964443624019623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,256,0.005616888817813661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,256,0.321889771355523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,128,0.011401777466138205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,64,0.0029093333416514923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1536,128,0.30926399760776097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1536,32,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,64,0.011383111278216044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,65536,0.02825688984658983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1536,32,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,65536,0.0855120023091634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,65536,0.34901067945692277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,16384,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,65536,0.05760622024536133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1536,128,0.005607111172543631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,16384,0.3521840042538113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,16384,0.016527111331621807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,12288,0.010544000400437249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,12288,0.02232088810867733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,12288,0.34293778737386066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,12288,0.013751111096805997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,10240,0.01071733319097095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,10240,0.020647999313142564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,16384,0.028519110547171697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,10240,0.336225774553087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,10240,0.012448888685968188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,8192,0.012696888711717395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,8192,0.01815733313560486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,8192,0.010991111397743225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,7168,0.011719110939237805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,8192,0.333912001715766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,7168,0.016930666234758165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,7168,0.010381333529949188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,7168,0.3381955623626709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,6144,0.016247999336984422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,6144,0.009768000079525841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,6144,0.32284800211588544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,5120,0.009659555223253038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,5120,0.015200000670221118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,5120,0.3379964563581679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,5120,0.008988444175985124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,4096,0.007664889097213745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,6144,0.011029333704047732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,4096,0.014475555883513557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,3584,0.006976000136799282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,4096,0.327111111746894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,4096,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,3584,0.014119999276267158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,3584,0.007991111112965478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,3072,0.0063484443558586975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,3584,0.31867021984524196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,3072,0.013790222505728403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,2560,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,3072,0.007669332954618666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,3072,0.3362729019588894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,2560,0.013472889032628802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,2048,0.006151999864313338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,2560,0.007275555696752336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,2560,0.3329182200961643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,2048,0.013125333521101209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,1536,0.005255110975768831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,2048,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,2048,0.32484979099697536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,1536,0.012726222475369772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,1024,0.004600888739029567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,1536,0.006622222148709827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,1536,0.33671198950873477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,1024,0.012411555482281579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,768,0.004192000048028098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,1024,0.006291555447710886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,1024,0.32013511657714844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,768,0.012152888708644442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,512,0.003548444559176763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,768,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,768,0.3324506547715929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,512,0.01205599970287747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,256,0.003552888830502828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,512,0.006010666489601135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,512,0.3267093234592014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,256,0.011533333195580376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,128,0.003160888950030009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,256,0.3306773238711887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,256,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,128,0.011700444751315646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,64,0.0029200000895394217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,1024,32,0.0032471112079090546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,1024,128,0.29542220963372123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,64,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,65536,0.02267733381854163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,1024,32,0.012127111355463663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,65536,0.08339377906587388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,16384,0.01553066737122006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,65536,0.3177644411722819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,65536,0.05765599674648709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,16384,0.02680177821053399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,16384,0.31661423047383624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,16384,0.01648622254530589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,12288,0.013447110851605734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,12288,0.02164266672399309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,1024,128,0.005589333259397083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,10240,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,12288,0.3127137819925944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,12288,0.013575110998418597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,10240,0.01995466649532318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,10240,0.012485333614879184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,8192,0.010460444622569613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,10240,0.319781329896715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,8192,0.01793511046303643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,7168,0.009727111293209923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,8192,0.010712888505723743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,8192,0.3408799966176351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,7168,0.016917333006858826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,7168,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,6144,0.008939555949634975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,7168,0.31938666767544216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,6144,0.016171556380059984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,6144,0.009663999908500248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,5120,0.008691555923885768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,6144,0.3458133273654514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,5120,0.015210666590266757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,4096,0.007657777931955125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,5120,0.009069333473841349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,5120,0.34129243426852757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,4096,0.014527110589875115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,3584,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,4096,0.343722661336263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,3584,0.014129777749379476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,3584,0.007655111451943715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,3584,0.3333288828531901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,3072,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,3072,0.01387466655837165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,3072,0.336641788482666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,3072,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,2560,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,4096,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,2560,0.01348088930050532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,2048,0.005294222384691238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,2560,0.32626401053534615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,2560,0.007259555160999298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,2048,0.013145777914259167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,1536,0.004553777890072929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,2048,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,2048,0.322181330786811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,1536,0.012800888882742988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,1024,0.003955555458863576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,1536,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,1536,0.3216053379906548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,1024,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,768,0.003640888879696528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,1024,0.00627377794848548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,1024,0.3234337700737847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,768,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,512,0.0035822221802340615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,768,0.005995555470387141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,768,0.3293520079718696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,512,0.01205599970287747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,256,0.0032302221904198327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,512,0.005656888915432825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,512,0.3249564435746935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,256,0.011776000261306763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,128,0.0028693332440323303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,256,0.005574222240183089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,256,0.32428889804416233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,128,0.01148888882663515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,64,0.002805333377586471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,768,128,0.3033084339565701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,768,32,0.002834666727317704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,64,0.011720000041855706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,65536,0.017836444907718234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,768,32,0.011423110961914062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,65536,0.08071644438637628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,65536,0.3329671223958333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,65536,0.056464003192053906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,16384,0.012426666915416718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,16384,0.0258879992696974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,12288,0.010729778144094678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,768,128,0.005609777652555042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,16384,0.01660711069901784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,16384,0.3181848790910509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,10240,0.009902222288979424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,12288,0.021685333715544805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,12288,0.013835555149449242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,12288,0.3238115575578478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,10240,0.019627556204795837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,8192,0.008678221868144141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,10240,0.012479999827014076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,8192,0.01792711185084449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,7168,0.008011555506123437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,8192,0.3146044413248698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,8192,0.011039111349317761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,7168,0.016826666063732572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,6144,0.007655111451943715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,7168,0.010406222608354356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,7168,0.3217742178175184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,6144,0.015824889143308003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,6144,0.009707555174827576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,6144,0.3720213307274713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,5120,0.00797155582242542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,5120,0.015459555718633862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,5120,0.3336302174462213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,4096,0.007631999750932057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,5120,0.009339555270142024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,4096,0.014389332797792224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,4096,0.4191253450181749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,3584,0.0069324444565508105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,3584,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,10240,0.34618043899536133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,3584,0.3258053461710612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,3584,0.008247999681366814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,3072,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,3072,0.013820444544156393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,3072,0.3307235505845812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,3072,0.007611555357774098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,2560,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,2560,0.01347644461525811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,2560,0.3167128827836779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,2560,0.0073520003093613525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,2048,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,2048,0.013122667041089801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,2048,0.3112764358520508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,1536,0.004907555464241239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,2048,0.006924444602595435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,4096,0.007983110845088959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,1536,0.33896711137559676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,1536,0.006549333532651265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,1024,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,1024,0.012157333393891653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,1024,0.33067109849717885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,1024,0.006258666515350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,768,0.003726222034957674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,768,0.011790222591824003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,768,0.3261253303951687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,768,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,512,0.003549333247873518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,1536,0.012847111456924014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,512,0.011768889096048145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,512,0.005647111270162795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,256,0.0032400000426504347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,512,0.32318488756815594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,256,0.011716444459226398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,128,0.003168888803985384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,256,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,256,0.3327351146274143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,128,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,64,0.00286311118139161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,512,128,0.005582222094138463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,512,32,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,512,128,0.31051111221313477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,64,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,512,32,0.01146044416560067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,65536,0.07854044437408447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,65536,0.3347253269619412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,65536,0.04998222324583265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,16384,0.009679999616410997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,16384,0.02601066728432973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,16384,0.33283999231126576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,12288,0.009706666900051964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,65536,0.01736088924937778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,12288,0.021371554997232225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,12288,0.013423110875818463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,12288,0.30848267343309194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,10240,0.009169778062237633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,10240,0.01957155598534478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,10240,0.012423111332787408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,8192,0.008265777594513362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,10240,0.30929777357313365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,8192,0.017906667457686532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,7168,0.007863110966152614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,16384,0.016490666402710807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,8192,0.33387290106879336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,7168,0.016846223009957206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,6144,0.00831911133395301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,7168,0.010074666804737514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,7168,0.29818667305840385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,6144,0.009662222531106737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,5120,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,6144,0.32013244099087185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,5120,0.015457777513398064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,5120,0.32886844211154515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,5120,0.009232000344329411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,4096,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,8192,0.011006222003036074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,4096,0.01477777792347802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,4096,0.008354666332403818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,3584,0.0069395556218094295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,4096,0.3108782238430447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,3584,0.014119999276267158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,3072,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,3584,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,6144,0.015973332855436537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,3584,0.31960444980197483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,2560,0.006311999840868845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,3072,0.013832888669437833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,3072,0.007596444752481248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,2560,0.013430222041077085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,3072,0.3355066776275635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,2048,0.004941333499219683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,2560,0.007516444557242923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,2560,0.3399662176767985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,2048,0.012829333543777466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,1536,0.00425866660144594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,2048,0.00700444479783376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,2048,0.32946133613586426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,1536,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,1024,0.003850666599141227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,1536,0.006599999964237213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,1536,0.33887378374735516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,1024,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,768,0.0035084444615576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,1024,0.006304889089531369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,1024,0.3278248839908176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,768,0.005936000082227919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,768,0.34267555342780215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,512,0.003536000019974179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,512,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,512,0.3417199982537164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,256,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,256,0.011719110939237805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,768,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,256,0.005655111124118169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,128,0.0028311111446883944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,128,0.011389333340856763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,512,0.005972444597217772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,128,0.31082222196790904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,256,128,0.005621333503060871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,64,0.0028773333049482773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,256,32,0.002855111120475663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,64,0.011682666838169098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,65536,0.014789332946141561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,256,256,0.3357848856184218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,65536,0.07598666350046794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,65536,0.048475556903415255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,65536,0.3315724531809489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,16384,0.009337777892748514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,16384,0.026131555438041687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,16384,0.352126227484809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,16384,0.01651822196112739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,12288,0.008299555215570662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,256,32,0.011423110961914062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,12288,0.021631111701329548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,10240,0.008691555923885768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,12288,0.013456888496875763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,12288,0.33194133970472545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,10240,0.019938665959570143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,8192,0.007970666719807519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,10240,0.012409778104888068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,10240,0.3238453335232205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,8192,0.018000889155599806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,7168,0.007963555554548899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,8192,0.010749333434634738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,8192,0.3294142352210151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,7168,0.016859556237856548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,6144,0.007621333003044128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,7168,0.010334221853150262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,7168,0.32448800404866535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,6144,0.01622577839427524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,5120,0.007017778025733099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,6144,0.009689778089523315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,6144,0.3343404399024115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,5120,0.01550222271018558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,4096,0.006592888798978593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,5120,0.3312648932139079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,4096,0.014671110444598727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,4096,0.00812266684240765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,4096,0.32036177317301434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,3584,0.014434667097197639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,3584,0.3120613363054064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,3584,0.00797155582242542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,3072,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,3072,0.01384888920519087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,5120,0.00904177791542477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,2560,0.0063164445261160536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,3072,0.31942489412095815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,3072,0.007309333317809635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,2560,0.01351199961370892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,2048,0.006220444622966979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,2560,0.3147786723242866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,3584,0.0064328888224230866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,2048,0.01313866674900055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,1536,0.004604444321658877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,2048,0.007028444773621029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,2048,0.3116106722089979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,1536,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,1024,0.003919110943873723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,1536,0.006607111129495833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,1024,0.012347555822796293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,1536,0.4487751324971517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,1024,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,768,0.0035422220826148987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,1024,0.3141644530826145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,2560,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,768,0.012108444339699216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,512,0.0034888887570963967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,768,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,512,0.012085333466529846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,768,0.4199368953704834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,256,0.0029502221279674103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,512,0.3033822112613254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,512,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,256,0.01167377746767468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,128,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,256,0.005650666852792104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,256,0.30876978238423664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,128,0.011363555159833698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,64,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,80,128,128,0.005555555638339784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,128,32,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,64,0.011397333608733283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,65536,0.013807111316257052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,128,32,0.01182933317290412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,16384,0.00832177781396442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,65536,0.07410311036639743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,12288,0.007659555309348636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,16384,0.025786666406525507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,10240,0.007148444652557373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,12288,0.021324444148275588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,8192,0.006637333167923822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,10240,0.01960088809331258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,7168,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,8192,0.017647999856207106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,80,128,128,0.3123511208428277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,6144,0.006319111006127463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,7168,0.016923555069499545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,5120,0.005990222096443176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,4096,0.005991999887757831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,6144,0.015863999724388123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,5120,0.015499555402331881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,3584,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,4096,0.014498665928840637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,3072,0.006228444476922353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,2560,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,3584,0.014194667339324951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,2048,0.00629333323902554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,3072,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,2560,0.013437333206335703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,2048,0.013167111410035027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,1024,0.003855111284388436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,1536,0.012493333054913415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,768,0.003842666745185852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,1024,0.012375999655988483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,512,0.0032391111469931076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,768,0.012132444315486483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,256,0.0031946665710873077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,512,0.012068444656001197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,128,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,256,0.011575111084514193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,64,0.0028071111689011254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,128,0.011720000041855706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,32,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,64,0.011418666276666852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,64,1536,0.00452888881166776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,65536,0.013794666363133324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,16384,0.00832622249921163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,64,32,0.012058667010731168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,12288,0.007657777931955125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,65536,0.07248355282677545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,16384,0.025762667258580525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,10240,0.007387555307812161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,8192,0.006609777609507243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,12288,0.021319111188252766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,7168,0.006635555376609166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,10240,0.0196106665664249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,6144,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,8192,0.017845332622528076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,5120,0.005958222266700532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,7168,0.01683644453684489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,4096,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,6144,0.01613333324591319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,5120,0.01514844430817498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,4096,0.014634667171372307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,3072,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,3584,0.014132445057233175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,2560,0.006566222343179915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,3072,0.013829333086808523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,2048,0.006280000011126201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,2560,0.013418667018413544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,1536,0.004582222137186262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,2048,0.013397333522637686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,1024,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,1536,0.012549333274364471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,768,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,1024,0.01240711079703437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,512,0.0035146665241983202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,768,0.012028444144460889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,256,0.0029102222373088202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,3584,0.0063635553750726916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,128,0.0028906667398081887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,512,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,64,0.002850666642189026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,256,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,80,32,32,0.0027102221631341507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,128,0.011424889167149862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,64,0.01166044423977534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,80,32,32,0.011717333561844297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,16384,0.34969600041707355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,16384,0.1694177786509196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,16384,0.1466719971762763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,12288,0.262442668279012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,12288,0.13260532750023737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,16384,0.3372355567084418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,12288,0.10575644175211589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,10240,0.21646666526794434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,10240,0.11147199736701117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,12288,0.35532087749905056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,8192,0.17979021867116293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,10240,0.10435733530256484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,8192,0.09608978033065796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,10240,0.33741243680318195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,8192,0.07794666952557035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,7168,0.15287289354536268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,7168,0.08278222216500176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,8192,0.3827333450317383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,6144,0.1350195540322198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,7168,0.07344710826873779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,7168,0.3175528844197591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,6144,0.07277955611546834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,5120,0.11039821969138251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,6144,0.058550225363837353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,5120,0.06298666530185275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,6144,0.33307822545369464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,4096,0.09455911318461101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,5120,0.04850044515397814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,5120,0.32682845327589244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,4096,0.04136088821623061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,3584,0.08024089203940497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,3584,0.04880977670351664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,4096,0.05919022030300564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,3584,0.3408666716681586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,3584,0.03775999943415324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,3072,0.06917688581678602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,4096,0.34615643819173175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,3072,0.04415377643373278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,2560,0.05901155869166056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,3072,0.03239733311865065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,2560,0.039325333303875394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,2048,0.049341334236992725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,3072,0.3279457887013753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,2560,0.029548443026012842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,2048,0.03441155619091458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,2048,0.023401778605249193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,1536,0.03784977727466159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,2560,0.4827466540866428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,2048,0.3203759988149007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,1024,0.027299554811583623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,1536,0.01921422282854716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,1024,0.023457777169015672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,1024,0.015368888775507608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,768,0.02123822271823883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,1024,0.41484978463914657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,768,0.019460444649060566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,768,0.013632000320487551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,512,0.014559111661381192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,768,0.3895706600613064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,1536,0.029330667522218492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,512,0.017214222086800467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,256,0.009016888837019602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,512,0.01164177805185318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,1536,0.33558400472005206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,512,0.3656364546881781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,128,0.006751110984219446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,256,0.015589333242840238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,256,0.010296000374688042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,256,0.32853955692715114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,128,0.014459555347760519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,64,0.00453333349691497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,65536,32,0.004548444516128964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,65536,128,0.30934045049879283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,64,0.015246222416559855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,65536,32,0.018230222993426852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,65536,0.32652534378899467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,65536,0.1935280031628079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,16384,0.08039733436372545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,16384,0.05856800079345703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,65536,128,0.010337777435779572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,65536,0.17338400416904023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,12288,0.06198755237791273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,16384,0.04615822103288439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,12288,0.04772177669737074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,16384,0.3497164514329698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,10240,0.05312444435225593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,12288,0.03839110996988084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,65536,0.3382559882269965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,12288,0.32359645101759166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,10240,0.04136711027887132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,8192,0.04339822133382162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,8192,0.03775555557674832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,10240,0.03836799992455377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,7168,0.038113776180479265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,8192,0.026030222574869793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,7168,0.03277511066860623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,7168,0.02677333354949951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,6144,0.03456177645259433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,7168,0.3233111169603136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,6144,0.029861334297392104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,6144,0.34122223324245876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,10240,0.3151742352379693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,5120,0.028821332587136164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,8192,0.31222934193081325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,6144,0.020745777421527438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,5120,0.027502222193611994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,4096,0.023403555154800415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,5120,0.01808088852299584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,4096,0.026102221674389307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,5120,0.32030222151014537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,3584,0.021498666869269475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,4096,0.015457777513398064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,4096,0.3363351027170817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,3584,0.0217866665787167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,3072,0.018783110711309645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,3584,0.014441778262456259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,3584,0.34719112184312606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,3072,0.020422221885787118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,2560,0.016196444630622864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,3072,0.013187555803192986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,3072,0.3346284495459662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,2560,0.018908444378111098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,2048,0.014075555735164218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,2560,0.012826667063766055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,2560,0.3390799893273248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,2048,0.01774666706720988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,1536,0.01069777790043089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,2048,0.010786666638321347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,2048,0.3277440071105957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,1536,0.016507556041081745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,1024,0.008412444757090675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,1536,0.009352000223265754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,1536,0.3414755662282308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,1024,0.014473777678277759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,768,0.007273777491516537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,1024,0.00795911086930169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,1024,0.33174578348795575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,768,0.013184888495339287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,512,0.00545866663257281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,768,0.007391999993059371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,768,0.3428337838914659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,512,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,512,0.32969421810574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,256,0.004190222256713443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,256,0.012392000191741519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,256,0.33417243427700466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,256,0.006585777633719974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,128,0.0035777779089079965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,128,0.012072888513406118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,16384,128,0.31245244873894584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,512,0.012458666331238218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,64,0.003548444559176763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,16384,128,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,16384,32,0.0038559999730851916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,64,0.012785777449607849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,16384,32,0.01313866674900055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,65536,0.23338490062289766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,65536,0.17421510484483507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,16384,0.06454666455586751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,65536,0.1480684412850274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,16384,0.34491999944051105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,16384,0.05288355549176534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,12288,0.050676445166269936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,65536,0.3453928894466824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,16384,0.04024444354905023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,12288,0.04302311274740431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,10240,0.04417066772778829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,12288,0.350200891494751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,12288,0.03305688831541274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,10240,0.0377786656220754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,10240,0.02719377809100681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,8192,0.03678755627738105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,10240,0.3352168930901422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,8192,0.03388266762097677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,7168,0.03204977843496535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,8192,0.33295199606153697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,8192,0.02277155551645491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,7168,0.030006223254733618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,6144,0.02796444296836853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,7168,0.022064000368118286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,6144,0.02695377833313412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,7168,0.3223662111494276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,5120,0.023912888434198167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,6144,0.01852888862291972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,5120,0.02370311154259576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,5120,0.016593777471118502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,4096,0.02028888960679372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,6144,0.45721154742770725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,5120,0.3183724350399441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,3584,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,4096,0.023024888502226934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,4096,0.014468444718254937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,4096,0.32945601145426434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,3584,0.020648888415760465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,3072,0.015475556254386902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,3584,0.013680000272062091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,3584,0.3411208788553874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,3072,0.012717333104875354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,2560,0.013776889277829064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,3072,0.29770488209194607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,2560,0.01792888840039571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,2560,0.3136204348670112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,2560,0.011777777638700275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,2048,0.012587555580668978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,2048,0.01700533264213138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,2048,0.31055911382039386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,1536,0.009658666948477427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,3072,0.01936177743805779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,1536,0.015645333462291293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,1536,0.3083893458048503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,1536,0.008415111237102086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,1024,0.007340444458855524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,1024,0.013288000391589271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,1024,0.32514577441745335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,1024,0.007644444704055786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,768,0.006621333460013072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,768,0.012798222402731577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,2048,0.00965244405799442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,512,0.004569777597983678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,768,0.3260257773929172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,768,0.007274666594134436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,512,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,256,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,512,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,512,0.3144480122460259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,256,0.011799111134476133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,256,0.006468444648716185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,128,0.003552000141806073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,256,0.3359688917795817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,128,0.012072000238630505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,64,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,12288,128,0.006647111227114995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,12288,128,0.3094080024295383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,12288,32,0.0035982223020659555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,64,0.01276711126168569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,12288,32,0.012097777591811286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,65536,0.2091439962387085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,65536,0.1679999960793389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,16384,0.050966223080952965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,65536,0.14115377267201742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,16384,0.05132711264822218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,16384,0.03858844439188639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,12288,0.04079733292261759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,16384,0.29292888111538357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,65536,0.3403680059644911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,12288,0.041929778125551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,10240,0.03457422057787577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,12288,0.031155556440353394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,12288,0.3025368849436442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,10240,0.0370071099864112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,8192,0.02848177817132738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,10240,0.02702577743265364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,10240,0.32405956586201984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,8192,0.032615999380747475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,7168,0.025489777326583862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,8192,0.022180444664425317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,7168,0.028555555476082697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,7168,0.020908445119857788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,6144,0.02216088937388526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,7168,0.3263315624660916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,6144,0.0255395554833942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,6144,0.3177013397216797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,6144,0.017670222454600863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,5120,0.018931556079122756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,5120,0.023711111810472276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,8192,0.32046400176154244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,4096,0.016574222180578444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,5120,0.01612711118327247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,4096,0.022277333670192297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,4096,0.3246364328596327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,4096,0.01386666629049513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,3584,0.014101333088344999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,3584,0.020290666156344943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,3072,0.012787555654843649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,3584,0.3179244465298123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,3584,0.013105777402718862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,3072,0.019038221902317472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,2560,0.011542222566074796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,3072,0.01184088902340995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,5120,0.31492000155978733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,3072,0.3187893231709798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,2048,0.01069866700304879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,2560,0.010752889017264048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,2560,0.3221493297153049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,2048,0.016878222425778706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,1536,0.008227555288208855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,2048,0.009318222602208456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,2048,0.31465421782599556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,1536,0.014891554911931356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,1024,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,1536,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,1536,0.3254088825649685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,1024,0.013111111190583972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,768,0.004967999955018361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,1024,0.00738844441043006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,1024,0.30836711989508736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,768,0.012788444757461548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,512,0.004171555654870139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,768,0.006999111009968652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,768,0.3242231210072835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,512,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,512,0.006613333192136552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,512,0.32215823067559135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,256,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,256,0.3215600119696723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,2560,0.018186666899257235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,128,0.003544888976547453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,256,0.006303999986913469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,128,0.011810666984981961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,256,0.0038808890514903595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,10240,128,0.3059075673421224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,10240,128,0.006608888920810487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,64,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,10240,32,0.003559111307064692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,64,0.011504000259770287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,10240,32,0.012778666284349231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,65536,0.11635022693210179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,65536,0.07751466830571492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,16384,0.04693688948949178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,65536,0.3374657895829942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,16384,0.03871377640300327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,65536,0.17034667068057588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,12288,0.03222666515244378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,16384,0.02397955622937944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,16384,0.30451644791497123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,12288,0.03257688879966736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,10240,0.02793955471780565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,12288,0.01885333326127794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,12288,0.33487465646531844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,10240,0.028734223710166082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,8192,0.02278311053911845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,10240,0.0166666673289405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,10240,0.325034671359592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,8192,0.026352889007992212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,7168,0.020909332566791113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,8192,0.01534755527973175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,8192,0.3321884473164876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,7168,0.022153778208626643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,6144,0.01867733399073283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,7168,0.013415111435784234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,7168,0.31632712152269155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,6144,0.020632000433074105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,5120,0.01579466627703773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,6144,0.012448888685968188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,6144,0.3237333297729492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,5120,0.019327999817000497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,4096,0.01312888910373052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,5120,0.011404444773991903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,5120,0.3207804361979167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,4096,0.019374221563339233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,3584,0.013129777378506131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,4096,0.009744000103738572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,4096,0.3287893401251899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,3584,0.01740622189309862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,3072,0.011578666667143503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,3584,0.009283555878533257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,3072,0.016551110479566786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,3072,0.008362666600280339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,2560,0.01071111112833023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,3072,0.3201022148132324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,2560,0.015810666812790763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,2560,0.28237422307332355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,2560,0.008316444853941599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,2048,0.014546665880415173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,3584,0.333987553914388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,2048,0.007381333245171442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,2048,0.3267173237270779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,1536,0.008072888685597314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,1536,0.013763555222087435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,1024,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,1536,0.007084444165229797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,1536,0.31476977136400014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,1024,0.01274311128589842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,2048,0.009476444787449306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,1024,0.006342222293217977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,768,0.012479999827014076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,1024,0.4490764405992296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,512,0.003951999876234266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,768,0.32203200128343373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,768,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,512,0.012116444607575735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,256,0.003536888708670934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,512,0.3151866594950358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,256,0.011913778053389655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,256,0.31745955679151744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,128,0.0035662220584021676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,768,0.004541333350870344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,128,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,8192,128,0.30768177244398326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,128,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,64,0.0032222221295038858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,8192,32,0.0035182221068276297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,512,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,64,0.011683555940786997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,8192,32,0.012061333490742577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,65536,0.150090667936537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,65536,0.10953778028488159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,16384,0.04629955689112345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,8192,256,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,65536,0.07543110847473145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,16384,0.03639288743336996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,16384,0.02372177773051792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,65536,0.3383359909057617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,16384,0.37119200494554305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,12288,0.0366284449895223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,12288,0.017868444323539734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,12288,0.031123555368847314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,10240,0.025711999999152288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,12288,0.3509644402398004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,10240,0.02715911136733161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,8192,0.021000888612535264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,10240,0.0162782222032547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,10240,0.3344746695624457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,8192,0.024088889360427856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,7168,0.01904088921017117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,8192,0.014816888504558139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,8192,0.34960444768269855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,7168,0.021327111456129286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,6144,0.017136888371573556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,7168,0.013082666529549493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,7168,0.33498311042785645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,6144,0.01997422178586324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,5120,0.014271999398867289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,6144,0.012125333150227865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,6144,0.3452186584472656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,5120,0.01901422275437249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,4096,0.01295555548535453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,5120,0.011115555961926779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,5120,0.3297191196017795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,4096,0.018592889110247295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,3584,0.011551111108726926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,4096,0.009316444396972656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,4096,0.34351465437147355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,3584,0.01716088917520311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,3072,0.01105955574247572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,3584,0.008984000318580205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,3584,0.3282719982994927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,3072,0.01626488897535536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,2560,0.010011555420027839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,3072,0.008348444269763099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,3072,0.3378329012129042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,2560,0.015196444259749519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,2048,0.009368889033794403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,2560,0.007629333270920648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,2560,0.3303137885199653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,2048,0.014107555150985718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,2048,0.007268444531493717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,1536,0.007366221812036302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,2048,0.3368871212005615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,1536,0.013410666750537025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,1024,0.0052168890833854675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,1536,0.00675644435816341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,1536,0.3311937650044759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,1024,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,768,0.004560889055331548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,1024,0.00628977765639623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,1024,0.33886310789320206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,768,0.012472888661755456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,512,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,768,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,768,0.32736976941426593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,512,0.012057777908113269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,256,0.003560888684458203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,512,0.005969777703285217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,512,0.3274639977349175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,256,0.011774222056070963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,128,0.003510222252872255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,256,0.0057653333577844836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,128,0.01148000028398302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,7168,128,0.005790222022268508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,128,0.30040356847974986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,64,0.0032257777121331957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,7168,32,0.003547555456558863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,64,0.012072000238630505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,7168,32,0.011752888560295105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,65536,0.11930755774180095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,65536,0.10193066464530097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,7168,256,0.3291955524020725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,16384,0.03787733448876275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,65536,0.07288444704479642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,16384,0.034677333301968045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,65536,0.3434444533454047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,12288,0.031015111340416804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,16384,0.022127999199761286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,16384,0.3322559992472331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,12288,0.02976177798377143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,12288,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,10240,0.027104000250498455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,12288,0.3247751129998101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,8192,0.02362222141689724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,10240,0.025999110605981614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,10240,0.01570044457912445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,8192,0.023066666391160753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,10240,0.3330390983157688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,7168,0.019897777173254225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,8192,0.014191110928853353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,8192,0.31278933419121635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,7168,0.0203484445810318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,6144,0.01856355534659492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,7168,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,7168,0.3302853372361925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,6144,0.019270223047998216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,5120,0.014899555179807874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,6144,0.011476444701353708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,6144,0.3259066740671794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,5120,0.01831022236082289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,4096,0.013800889253616333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,5120,0.010451555252075195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,5120,0.32865956094529897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,4096,0.01809244520134396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,3584,0.011738667057620155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,4096,0.009135999613338048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,4096,0.3372017807430691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,3584,0.01661777827474806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,3584,0.008436444732877944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,3584,0.33482223086886936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,3072,0.015820443630218506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,3072,0.008016000191370646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,3072,0.3355982303619385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,2560,0.014430221584108142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,3072,0.009679111341635386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,2560,0.007617777420414819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,2048,0.008649778034951951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,2560,0.008683555656009251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,2048,0.3126151031917996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,2048,0.007304000357786815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,2560,0.3337457709842258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,1536,0.00656355544924736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,1536,0.013103110922707452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,1536,0.33333688312106663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,1536,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,2048,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,1024,0.012432888978057437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,1024,0.30659111340840656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,1024,0.006295111030340195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,768,0.0041955556306574075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,768,0.01236355553070704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,768,0.31341422928704155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,512,0.0038906666967603895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,512,0.012085333466529846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,1024,0.005032000028424793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,512,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,256,0.0035013332962989807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,512,0.332407103644477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,256,0.011789333489206104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,256,0.005988444305128521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,256,0.3164337741004096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,128,0.0031662223239739737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,768,0.005956444475385878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,128,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,64,0.0031795555518733132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,6144,128,0.005910222315125995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,6144,128,0.3138604429033068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,6144,32,0.0032142222755485107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,64,0.01184177812602785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,65536,0.104239112801022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,6144,32,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,65536,0.09386310974756877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,16384,0.032610668076409235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,65536,0.07032977872424655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,16384,0.032348444064458214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,12288,0.026604443788528442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,65536,0.3408968978457981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,16384,0.02100444502300686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,16384,0.3346337742275662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,10240,0.023750222391552393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,12288,0.016954667038387723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,12288,0.027112000518374976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,10240,0.0232595553000768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,8192,0.020219556159443326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,10240,0.015358222855461968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,10240,0.32411466704474556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,8192,0.021597334080272253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,7168,0.018633777896563213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,8192,0.013735111388895245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,7168,0.019658666517999437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,8192,0.3969759941101074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,6144,0.01420177850458357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,7168,0.319538672765096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,7168,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,6144,0.018394667241308425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,6144,0.01108444482088089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,5120,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,12288,0.32209332784016925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,6144,0.3145528899298774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,4096,0.010959111154079437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,5120,0.01755466726091173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,5120,0.010068444742096795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,4096,0.01776444415251414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,5120,0.3440728982289632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,4096,0.008354666332403818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,4096,0.3342711130777995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,3584,0.01588088936275906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,3072,0.008957333034939235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,3584,0.008311111066076491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,3584,0.32250221570332843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,3072,0.014127999544143677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,2560,0.008117333054542542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,3072,0.007707555260923173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,3072,0.32203110059102374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,2560,0.013839110732078552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,2560,0.3292151027255588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,2560,0.007452444069915348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,2048,0.007667555577225155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,3584,0.009496889180607265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,2048,0.013099555340078143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,2048,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,1536,0.005246222019195557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,2048,0.33029156261020237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,1536,0.012772444221708508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,1024,0.004436444491147995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,1536,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,1536,0.33075467745463055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,1024,0.012627555264366997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,768,0.0042133331298828125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,1024,0.006239111224810283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,1024,0.3375306659274631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,768,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,512,0.003871110992299186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,768,0.00598933340774642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,768,0.3129182126786974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,512,0.012135999898115793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,256,0.003516444315512975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,512,0.005896889087226655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,512,0.3334542115529378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,256,0.01201244443655014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,128,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,256,0.005865777946180767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,256,0.31721067428588867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,128,0.011693333586057028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,64,0.0031582222630580268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,5120,128,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,5120,32,0.0031591111587153543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,5120,128,0.30573511123657227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,64,0.01181688904762268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,65536,0.08443999952740139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,5120,32,0.011086222198274402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,65536,0.09195733070373535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,16384,0.027140445179409448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,65536,0.06769777668846978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,16384,0.03052355514632331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,65536,0.3361990981631809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,12288,0.021891555852360193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,16384,0.020305777589480083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,12288,0.025437333517604407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,12288,0.33227822515699595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,12288,0.016599110431141324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,10240,0.019175110591782462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,10240,0.022293332550260756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,10240,0.33461777369181317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,10240,0.014860444598727755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,8192,0.015597333510716757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,16384,0.3305022186703152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,8192,0.020807999703619216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,7168,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,8192,0.3302080101437039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,8192,0.012897777888509961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,7168,0.018806222412321303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,6144,0.013542222479979197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,7168,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,7168,0.32255644268459743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,6144,0.018013333280881245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,5120,0.011086222198274402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,6144,0.01069777790043089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,6144,0.3262382348378499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,5120,0.017231111725171406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,4096,0.010259555445777046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,5120,0.009333333207501305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,5120,0.33162132898966473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,4096,0.015842666228612263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,3584,0.008798222574922774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,4096,0.008328888979223039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,4096,0.3205733299255371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,3584,0.014578666951921252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,3072,0.008028444316652086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,3584,0.007993777592976889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,3584,0.32052800390455455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,3072,0.014160888062583076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,2560,0.007663110891977946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,3072,0.007682667010360294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,3072,0.3173706796434191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,2560,0.013776889277829064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,2048,0.006573333508438534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,2560,0.329861323038737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,2048,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,2048,0.0069573331210348345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,2048,0.3215920130411784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,1536,0.005210666606823604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,1536,0.01274222218328052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,1024,0.004226666771703296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,1536,0.3303271134694417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,1536,0.006766222417354584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,2560,0.00738666703303655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,768,0.0038808890514903595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,1024,0.33046044243706596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,1024,0.00666933341158761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,768,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,512,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,768,0.006033777776691649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,768,0.3177155653635661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,512,0.011782222323947482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,512,0.006010666489601135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,512,0.403730657365587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,256,0.003173333282272021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,1024,0.01254755589697096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,256,0.011539555258221097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,128,0.0031991110493739447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,256,0.005576888720194499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,256,0.33064267370435924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,128,0.011765333513418833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,4096,128,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,64,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,4096,32,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,4096,128,0.31807467672559947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,64,0.010822222464614444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,65536,0.07414755556318495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,4096,32,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,65536,0.08853155374526978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,16384,0.02511377798186408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,65536,0.06690222024917603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,65536,0.33497601085238987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,16384,0.028982222080230713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,12288,0.01998844411638048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,16384,0.37094489733378094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,16384,0.020251555575264823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,12288,0.02424088948302799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,10240,0.018684445155991446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,12288,0.016365332735909354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,12288,0.3542017671797011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,10240,0.021635555558734473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,8192,0.01513777838812934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,10240,0.014821334017647637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,10240,0.3370853265126546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,8192,0.020079111059506733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,8192,0.012463111016485425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,7168,0.014125333891974555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,8192,0.35124890009562176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,7168,0.018696000178654987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,6144,0.012484444512261285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,7168,0.011404444773991903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,7168,0.326200008392334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,6144,0.017822222577200997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,5120,0.010579555398888059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,6144,0.010629333555698395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,6144,0.33531912167867023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,5120,0.01665155589580536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,4096,0.009859555297427708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,5120,0.009315555294354757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,5120,0.33342578676011825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,4096,0.015279110934999255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,3584,0.008712000317043727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,4096,0.008303999900817871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,4096,0.33348266283671063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,3584,0.01441333360142178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,3072,0.00796977761718962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,3584,0.008009778128729926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,3072,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,3072,0.007634667058785756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,3072,0.2993626594543457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,2560,0.0069217777086628815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,2560,0.013397333522637686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,2560,0.3223280111948649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,2560,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,2048,0.005886222339338726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,2048,0.013068444199032254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,3584,0.3031502299838596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,2048,0.32029512193467885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,1536,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,2048,0.006959111326270633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,1536,0.012820444173283048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,1536,0.006647111227114995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,1024,0.0041911109454101985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,1536,0.30754134390089244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,1024,0.012443555725945367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,768,0.004034666551484002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,1024,0.006301333506902059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,1024,0.32270754708184135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,768,0.012155555188655853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,512,0.00351555562681622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,768,0.005889777921968036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,768,0.3222222328186035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,512,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,256,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,512,0.005946666830115848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,512,0.33170223236083984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,256,0.011683555940786997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,128,0.0029760001020299066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,256,0.005566222386227713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,256,0.31527378824022084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,64,0.0031715554909573663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3584,128,0.005575110928879843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3584,32,0.0031822222388452957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,64,0.010833777487277985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3584,128,0.43602667914496523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,32,0.011769777370823754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,65536,0.06420533524619208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,65536,0.08771910932328965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,16384,0.023476445012622412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,65536,0.06598844130833943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,65536,0.3338613245222304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,16384,0.027849776877297297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,16384,0.3286728858947754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3584,128,0.011368888947698804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,12288,0.020231111182106864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,16384,0.0196435550848643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,12288,0.02421422302722931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,10240,0.01755199995305803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,12288,0.016041777200169034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,12288,0.32682132720947266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,10240,0.02130577796035343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,8192,0.015427556302812366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,10240,0.014318222800890604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,10240,0.3688666555616591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,8192,0.019926221834288705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,7168,0.013887999786270989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,8192,0.01110666659143236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,8192,0.32366845342848033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,7168,0.018248000078731112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,6144,0.01200977795653873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,7168,0.01109688894616233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,7168,0.3492773373921712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,6144,0.017447999782032438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,5120,0.010324444207880232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,6144,0.009679999616410997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,6144,0.32501599523756236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,5120,0.015949333707491558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,4096,0.009336888790130615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,5120,0.009023110899660323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,5120,0.3416817718082004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,4096,0.014854222536087036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,3584,0.008350222474998897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,4096,0.008375110725561777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,4096,0.3243244489034017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,3584,0.014192000031471252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,3072,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,3584,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,3584,0.34094399876064724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,3072,0.013737777868906656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,2560,0.006585777633719974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,3072,0.0076453329788313965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,3072,0.3172310988108317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,2560,0.013434666726324292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,2048,0.005573333137565189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,2560,0.007291555404663086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,2560,0.33861509958902997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,2048,0.01313066648112403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,1536,0.004954666727119022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,2048,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,2048,0.33079467879401314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,1536,0.012764444781674279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,1024,0.004239999999602635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,1536,0.006675555474228329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,1536,0.3230915599399143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,1024,0.012381333443853589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,768,0.003914666672547658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,1024,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,1024,0.3295911153157552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,768,0.012166221936543783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,512,0.0038568890757030914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,768,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,512,0.01204711116022534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,768,0.34345777829488117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,256,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,512,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,512,0.3222035566965739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,256,0.01147555559873581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,128,0.0029306666304667792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,256,0.005566222386227713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,256,0.29778043429056805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,128,0.011676444775528379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,64,0.0028844444702068963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,3072,128,0.005604444278611078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,3072,128,0.28969155417548287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,3072,32,0.0032400000426504347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,64,0.011592888997660743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,65536,0.05642133288913303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,3072,32,0.010794666906197866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,65536,0.08352000183529323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,16384,0.019010666343900893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,65536,0.06474400228924222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,65536,0.3320862187279595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,16384,0.027062222361564636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,12288,0.01826844447188907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,16384,0.3283369011349148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,16384,0.019567999574873183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,12288,0.023423999547958374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,10240,0.016524444023768108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,12288,0.015543111496501498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,12288,0.3214248816172282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,10240,0.021047110358874004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,8192,0.014488000008794995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,10240,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,10240,0.33324265480041504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,8192,0.01957244508796268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,7168,0.013152889079517789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,8192,0.011429333024554782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,8192,0.31659822993808323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,7168,0.01795733306143019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,6144,0.012491555677519904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,7168,0.011191999746693505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,7168,0.3361084461212158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,6144,0.01679111189312405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,5120,0.00981244444847107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,6144,0.00978577799267239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,6144,0.3296977678934733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,5120,0.015519110692871941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,4096,0.00923822240697013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,5120,0.009321778184837764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,4096,0.014495111174053617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,5120,0.4593022134568956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,3584,0.008070222205585903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,4096,0.3318471113840739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,3584,0.014538667268223233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,3584,0.32815466986762154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,3072,0.006959111326270633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,3584,0.008048888709810045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,3072,0.013470222552617392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,3072,0.3888168864780002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,2560,0.006258666515350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,4096,0.008345777789751688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,2560,0.013491555220550962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,2560,0.007317333585686154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,2560,0.34470399220784503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,2048,0.005496888938877318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,2048,0.013125333521101209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,2048,0.3638924492730035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,2048,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,1536,0.00488800017370118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,3072,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,1536,0.012814222110642327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,1536,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,1536,0.34169332186381024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,1024,0.012466666599114736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,1024,0.006340444501903322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,768,0.003925333420435588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,1024,0.3316284550560845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,768,0.012153777811262341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,768,0.329931550555759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,768,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,1024,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,512,0.012025777664449481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,512,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,256,0.0035173334181308746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,512,0.3379555543263753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,256,0.011791999969217511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,256,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,256,0.335024012459649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,128,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,512,0.003628444340493944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,128,0.011427555647161273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,64,0.0031786666562159858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2560,128,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2560,128,0.31755820910135907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2560,32,0.003169777699642711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,65536,0.04550933175616794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,32,0.011447110937701331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,65536,0.0832586685816447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,65536,0.06358844704098172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,16384,0.01531644496652815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,16384,0.026738666825824316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2560,64,0.011776888536082374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,16384,0.2981857723659939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,12288,0.013080000049538083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,16384,0.018968888454967074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,12288,0.02309688925743103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,12288,0.014800889624489678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,65536,0.33939554956224227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,10240,0.013360888593726687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,10240,0.012484444512261285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,8192,0.01296533313062456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,10240,0.30460800064934623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,8192,0.01866755551762051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,8192,0.316140439775255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,8192,0.011047110789351992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,7168,0.011071110765139261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,7168,0.01719111038578881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,7168,0.32625778516133624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,7168,0.010379555324713389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,6144,0.010432889064153036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,6144,0.016143111719025504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,10240,0.02090399960676829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,5120,0.010041777458455827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,6144,0.30959200859069824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,6144,0.009708444277445475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,5120,0.014801777071423002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,5120,0.008987555901209513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,4096,0.008482666479216682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,5120,0.3068853219350179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,4096,0.014452444182501899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,3584,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,4096,0.008330666356616551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,4096,0.3122017648484972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,12288,0.3471653461456299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,3584,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,3072,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,3584,0.007987555530336168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,3584,0.3257288932800293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,2560,0.005877333382765452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,3072,0.013792888985739814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,3072,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,3072,0.33732623524136013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,2560,0.013452444639470843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,2048,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,2560,0.32477156321207684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,2560,0.007327111230956183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,2048,0.01313066648112403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,1536,0.004574222283230888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,2048,0.00700444479783376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,2048,0.3204444514380561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,1536,0.012721777790122561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,1024,0.0041911109454101985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,1536,0.006617777877383762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,1536,0.3335786660512288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,1024,0.01238844460911221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,768,0.003843555433882607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,1024,0.006243555496136348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,1024,0.30449599689907497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,768,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,512,0.003563555578390757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,768,0.005967999911970562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,768,0.3279040124681261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,512,0.012135999898115793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,256,0.0033359999457995095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,512,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,512,0.3159528838263618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,256,0.011714666253990598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,128,0.002942222274012036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,256,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,256,0.3366559876335992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,128,0.01145600030819575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,64,0.0029279999434947968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,2048,128,0.005606222069925732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,2048,32,0.0028311111446883944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,2048,128,0.3022186756134033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,65536,0.03573688864707947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,64,0.011367999845080905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,2048,32,0.01180266671710544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,65536,0.08134133285946317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,16384,0.014698666003015308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,65536,0.06197066439522637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,65536,0.34255109892951113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,16384,0.026328000757429335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,12288,0.01332800007528729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,16384,0.33679911825391984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,12288,0.022555554906527202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,12288,0.32382043202718097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,10240,0.011422221859296164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,10240,0.020586666133668687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,10240,0.32860978444417316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,16384,0.018629334039158292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,8192,0.010458666417333813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,10240,0.012510221865442065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,8192,0.01723911033736335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,12288,0.013823111024167804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,7168,0.010087110930018956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,8192,0.011041777829329172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,8192,0.3427644570668538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,7168,0.0165928883685006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,7168,0.010344888601038191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,6144,0.010032888915803697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,7168,0.3394355509016249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,6144,0.016217778126398723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,5120,0.009004444711738164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,6144,0.00981244444847107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,6144,0.3323475519816081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,5120,0.015188443991872998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,4096,0.008016000191370646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,5120,0.008992888861232335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,5120,0.3357902103000217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,4096,0.014488000008794995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,3584,0.007267555428875818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,4096,0.008354666332403818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,4096,0.3297955460018582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,3584,0.013880000346236758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,3072,0.006631999793979857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,3584,0.3368488947550456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,3072,0.013767110804716745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,3072,0.00758133331934611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,3072,0.3318657875061035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,2560,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,2560,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,2560,0.3438275655110677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,2048,0.005470222069157495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,2560,0.007342221836249034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,2048,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,2048,0.007024889190991719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,3584,0.008060444560315874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,1536,0.004828444371620814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,1536,0.012438221938080259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,1536,0.006646222124497096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,1024,0.003915555361244413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,2048,0.4659342235989041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,1536,0.3203146722581652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,768,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,1024,0.012171555724408893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,1024,0.006362666686375936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,1024,0.3377813233269586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,768,0.01239377756913503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,512,0.0031991110493739447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,768,0.00600622221827507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,768,0.3325164318084717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,512,0.012061333490742577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,256,0.0032808888289663526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,512,0.005909333212508096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,512,0.34594400723775226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,256,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,128,0.003177777760558658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,256,0.00563466673096021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,256,0.33561865488688153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,128,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,64,0.0028862222615215513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1536,128,0.005649777750174205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1536,128,0.31269778145684135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1536,32,0.002855111120475663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,64,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,65536,0.0251511103577084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1536,32,0.011431111229790581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,65536,0.07839822106891207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,16384,0.012426666915416718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,65536,0.05768711037105984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,65536,0.34196800655788845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,16384,0.026009778181711834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,16384,0.016382222374280293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,12288,0.01035733355416192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,16384,0.3495493464999729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,12288,0.020954666866196528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,10240,0.010388444695207808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,12288,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,12288,0.3402675522698297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,10240,0.019302222463819716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,10240,0.012483555409643384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,8192,0.009205333060688442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,10240,0.33755021625094944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,8192,0.01760444376203749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,7168,0.009010666774378883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,8192,0.011008000208271874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,8192,0.3419671058654785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,7168,0.01648977730009291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,6144,0.008095111283991072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,7168,0.3469857904646132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,6144,0.015834665960735746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,6144,0.009670221971140968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,6144,0.33255288336012095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,5120,0.008353778057628209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,5120,0.014836443795098199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,5120,0.33864177597893613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,5120,0.009026666482289633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,4096,0.007637333538797166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,7168,0.010412444670995077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,4096,0.014184888866212634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,3584,0.007345777418878343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,4096,0.32739822069803876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,3584,0.013897778259383308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,3584,0.2944844563802083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,3584,0.008050666915045844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,3072,0.006350222147173352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,3072,0.013510222236315409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,3072,0.3100364473130968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,3072,0.007676444119877285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,2560,0.0058764442801475525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,2560,0.013199999928474426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,4096,0.008040889269775814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,2560,0.30698132514953613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,2048,0.0052871112194326185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,2560,0.007262222468852997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,2048,0.013099555340078143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,1536,0.00453599997692638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,2048,0.3463253445095486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,1536,0.012746666868527731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,1536,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,1024,0.004152889053026835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,1536,0.3238328827752007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,1024,0.01239999963177575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,1024,0.3242444462246365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,1024,0.00628355559375551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,768,0.003863999827040566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,2048,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,768,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,512,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,768,0.3237351046668159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,768,0.005965333431959152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,512,0.012062222593360476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,256,0.0031866667171319327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,512,0.3283884525299072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,256,0.011808888779746162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,128,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,256,0.005626666463083691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,128,0.011447110937701331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,256,0.46056000391642254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,64,0.002899555489420891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,128,0.005647999958859549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,1024,128,0.30307287640041775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,1024,32,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,65536,0.02401688860522376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,64,0.011133333047231039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,1024,32,0.011432888607184092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,65536,0.07675022549099393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,1024,512,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,16384,0.01163200040658315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,65536,0.057408001687791615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,16384,0.02476799984773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,65536,0.41004178259107804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,12288,0.00960533320903778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,16384,0.32452087932162815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,16384,0.016561778055297006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,12288,0.020696888367335003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,10240,0.010034666293197209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,12288,0.014162666267818876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,12288,0.32768887943691677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,10240,0.018911111685964797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,8192,0.008719111482302347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,10240,0.012411555482281579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,10240,0.3404551082187229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,8192,0.016891555653678048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,8192,0.010949333508809408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,7168,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,8192,0.33249245749579537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,7168,0.016220443778567843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,7168,0.0103493332862854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,6144,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,7168,0.3491697841220432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,6144,0.015896888242827523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,5120,0.008230222596062554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,6144,0.009679999616410997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,6144,0.33610044585333926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,5120,0.01514311134815216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,4096,0.007418666448858049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,5120,0.33668798870510525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,4096,0.014509333504570855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,4096,0.008027555214034187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,4096,0.33673599031236434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,3584,0.0075288886825243635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,3584,0.013814222481515674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,3584,0.3482186794281006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,3584,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,3072,0.006728000111050076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,5120,0.009032888544930352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,3072,0.013451555536852943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,2560,0.006639110959238476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,3072,0.007613333563009898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,3072,0.32901154624091256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,2560,0.013117333253224691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,2048,0.004916444420814514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,2560,0.007307555940416124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,2560,0.32760000228881836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,2048,0.01276711126168569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,1536,0.004552888787455029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,2048,0.0069297779765393995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,2048,0.3230862352583143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,1536,0.012493333054913415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,1536,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,1536,0.3263137870364719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,1024,0.012468444804350534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,1024,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,768,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,1024,0.32811999320983887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,768,0.012303111453851065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,768,0.337990230984158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,768,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,512,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,1024,0.0038533334930737815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,512,0.012072888513406118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,256,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,512,0.3353244463602702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,256,0.011770666473441653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,256,0.005580444302823808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,256,0.3270675606197781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,128,0.0028284444577164124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,128,0.011420444481902652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,768,128,0.3088302347395155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,64,0.002887111157178879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,64,0.011757333245542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,768,32,0.0028151110228565005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,512,0.005660444498062134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,65536,0.01776266594727834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,768,32,0.011735999749766456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,65536,0.07287022140291002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,16384,0.010491555763615502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,65536,0.05573066737916735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,65536,0.34088177151150173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,16384,0.0237137774626414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,12288,0.00902133352226681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,16384,0.016519111063745286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,768,128,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,12288,0.020247111717859905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,12288,0.013727111121018728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,16384,0.4427226649390326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,10240,0.009009777671760982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,12288,0.3300995561811659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,8192,0.008307555483447181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,10240,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,8192,0.017326222525702584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,10240,0.39455021752251523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,7168,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,8192,0.01105955574247572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,7168,0.016372443901167977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,7168,0.33228445053100586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,7168,0.010379555324713389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,6144,0.0073626670572492815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,6144,0.015677332878112793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,10240,0.019003555178642273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,6144,0.3260471026102702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,5120,0.008027555214034187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,6144,0.009719111025333405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,5120,0.01517688896920946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,8192,0.32540178298950195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,4096,0.007017778025733099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,5120,0.33565955691867405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,4096,0.014112888111008538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,4096,0.3278462092081706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,4096,0.007984889050324758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,3584,0.0069466663731469055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,3584,0.013790222505728403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,3584,0.339499553044637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,3584,0.007687999970383114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,3072,0.006752000086837345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,3072,0.013474666410022311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,5120,0.008969777988062965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,2560,0.006581333362393909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,3072,0.3171982235378689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,3072,0.007293333609898885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,2560,0.013481777575280933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,2048,0.006536888993448681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,2560,0.0073315559162033936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,2560,0.32002221213446724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,2048,0.01314755529165268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,1536,0.004543111142185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,2048,0.0069120000633928515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,2048,0.3368862205081516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,1536,0.01276177747382058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,1024,0.004201777693298128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,1536,0.006341333190600078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,1536,0.3262346585591634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,1024,0.012458666331238218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,768,0.0038737778862317405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,1024,0.00630044440428416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,1024,0.329133325152927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,768,0.012157333393891653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,512,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,768,0.005943111247486538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,768,0.33055022027757436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,512,0.012148444851239523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,256,0.0032177778581778207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,512,0.32729511790805393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,256,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,128,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,256,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,256,0.32615555657280815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,128,0.011526222030321756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,64,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,512,128,0.005656888915432825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,512,32,0.00282222218811512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,512,128,0.3005439970228407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,65536,0.014871999621391296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,32,0.01147911118136512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,65536,0.0704631143146091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,65536,0.345944881439209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,65536,0.04935822221967909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,16384,0.009716444545321995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,16384,0.023471999499532912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,16384,0.32975021998087567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,12288,0.00833333366447025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,16384,0.016238222519556682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,512,64,0.011734222372372946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,12288,0.02037244372897678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,10240,0.007677333222495184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,12288,0.013722666435771518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,12288,0.33618489901224774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,10240,0.018946665856573317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,8192,0.007280888656775157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,10240,0.012418666647540199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,10240,0.3359786669413249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,8192,0.017055999901559617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,7168,0.006981333096822103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,8192,0.010984888507260216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,8192,0.332294225692749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,7168,0.01622222198380364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,6144,0.00658044425977601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,7168,0.010110222631030613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,7168,0.3282284471723768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,5120,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,6144,0.015453333655993143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,6144,0.009769777456919352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,5120,0.015209777487648858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,4096,0.006919999917348226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,5120,0.32914667659335667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,5120,0.009090666969617208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,4096,0.014474666780895658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,3584,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,4096,0.008057778080304464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,4096,0.3203351232740614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,3584,0.014087110757827759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,3072,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,3584,0.007713778151406183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,3072,0.013722666435771518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,3584,0.4002835485670302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,2560,0.006690666493442323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,3072,0.007458666960398356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,6144,0.32101066907246906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,3072,0.3201199902428521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,2560,0.013178666432698568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,2560,0.007241778075695038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,2560,0.33668356471591526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,2048,0.012990222209029727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,2048,0.007003555695215861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,2048,0.32112622261047363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,1536,0.012830221818553077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,1536,0.3116613229115804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,2048,0.0063279999627007385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,1024,0.0038382220599386427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,1536,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,1024,0.012503110700183444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,1536,0.00452622233165635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,1024,0.33224532339307994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,768,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,768,0.012085333466529846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,768,0.3150888813866509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,768,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,512,0.012115555504957834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,512,0.32846400472852916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,1024,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,256,0.0031635556370019913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,512,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,256,0.011448000040319232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,512,0.0032017777363459268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,256,0.3296906683180067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,128,0.0028391112056043413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,256,0.005656888915432825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,128,0.011521778172916837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,256,128,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,64,0.0025351111673646504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,256,32,0.0025351111673646504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,256,128,0.29701956113179523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,65536,0.014375999569892883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,64,0.01148888882663515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,256,32,0.011439999772442712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,65536,0.07003021902508207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,16384,0.00832266691658232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,65536,0.04895377821392483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,65536,0.3468115594651964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,16384,0.023729777998394434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,12288,0.007679999702506595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,16384,0.01625066664483812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,16384,0.33076710171169704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,12288,0.020652444826232064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,10240,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,12288,0.350816011428833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,10240,0.018914666440751817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,10240,0.012400888734393649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,8192,0.0070062221752272705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,10240,0.3365768856472439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,8192,0.017236444685194228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,8192,0.33431821399264866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,8192,0.010750222537252637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,7168,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,12288,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,7168,0.016152888536453247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,6144,0.006583999842405319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,7168,0.32969866858588326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,7168,0.010355555348926121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,6144,0.015568888849682279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,5120,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,6144,0.3206640084584554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,5120,0.014836443795098199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,5120,0.00903733323017756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,4096,0.005969777703285217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,5120,0.3249555428822835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,4096,0.014375999569892883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,4096,0.33434221479627824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,3584,0.0063271112740039825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,6144,0.009672889278994666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,3584,0.013835555149449242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,3072,0.005985777825117111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,3584,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,3072,0.013767999907334646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,3584,0.48076799180772567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,2560,0.006353777729802662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,3072,0.0076044441925154785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,2560,0.013136000268989138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,2560,0.3349733352661133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,4096,0.008078222473462423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,2048,0.005983110931184556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,2560,0.007060444189442529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,2048,0.013060444758998023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,1536,0.004608889006906086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,2048,0.006646222124497096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,2048,0.32141955693562824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,1536,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,1024,0.003928889003064898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,1536,0.006577777779764599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,1536,0.3740639951494005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,3072,0.31768621338738334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,1024,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,768,0.0035706667436493766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,1024,0.006117333140638139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,1024,0.3216053379906548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,768,0.012145777543385824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,512,0.0034764446318149567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,768,0.0057591112951437635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,768,0.3177306652069092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,512,0.011821333732869891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,256,0.0031973332580592898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,512,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,512,0.32949333720737034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,256,0.011735111474990845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,128,0.0029377777957253983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,256,0.005615111026499007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,256,0.3181777795155843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,128,0.011422221859296164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,64,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,64,128,128,0.005589333259397083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,64,128,128,0.30362844467163086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,128,32,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,64,0.011162666810883416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,65536,0.011092444260915121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,16384,0.007290666302045186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,128,32,0.011408000356621213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,12288,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,65536,0.0662942197587755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,16384,0.024030221833123103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,12288,0.02033244404527876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,8192,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,10240,0.019111999207072787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,7168,0.00666844430896971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,8192,0.01716444392999013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,6144,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,7168,0.016527111331621807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,5120,0.006615110983451207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,6144,0.015815110670195687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,4096,0.005972444597217772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,5120,0.015160888433456421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,3584,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,4096,0.014185777968830533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,3072,0.005975111077229182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,10240,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,3584,0.014173333843549093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,2560,0.006272888845867581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,2048,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,3072,0.013762666947311826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,2560,0.013427555561065674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,1536,0.004270222038030624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,1024,0.003830222205983268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,2048,0.012900444368521372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,768,0.0036115555299652945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,1536,0.012806221842765808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,512,0.00349866681628757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,1024,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,256,0.0029271110478374693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,768,0.012113778127564324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,128,0.0028497777465316984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,512,0.012084444363911947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,64,0.0028497777465316984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,256,0.011427555647161273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,64,32,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,64,0.011556444068749746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,65536,0.011032000184059143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,32,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,16384,0.007338666253619724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,65536,0.06707644462585449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,12288,0.006583999842405319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,16384,0.023744889431529578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,10240,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,12288,0.020693333612547982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,8192,0.006917333437336816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,10240,0.018761777215533786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,7168,0.006585777633719974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,8192,0.017549332645204332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,6144,0.006588444527652528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,7168,0.016347555650605094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,5120,0.0063235556913746735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,6144,0.015819556183285184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,4096,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,5120,0.015096000499195524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,3584,0.006240889016124938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,4096,0.014456889695591398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,3072,0.005899555567238066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,3584,0.013904889424641928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,2560,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,3072,0.013792888985739814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,2048,0.005987555616431766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,2560,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,1536,0.004593777573770947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,2048,0.013108444710572561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,1024,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,1536,0.012727110750145383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,768,0.00360977773865064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,1024,0.012416000167528788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,512,0.003346666693687439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,768,0.012111999922328524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,256,0.002962666667169995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,512,0.012094222009181976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,128,0.002847111059559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,256,0.011587555209795633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,128,0.011618666350841522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,64,32,32,0.002560888934466574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,64,0.011454222102959951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,32,32,0.011436444189813403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,64,64,128,0.011428444749779172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,16384,0.34802844789293075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,16384,0.16700977749294707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,12288,0.2591128879123264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,16384,0.15328889422946507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,12288,0.13067021634843615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,12288,0.10450488991207546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,16384,0.35156978501213926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,10240,0.21463378270467123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,10240,0.11015111207962036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,12288,0.35103999243842227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,8192,0.17950222227308485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,10240,0.10023288594351874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,8192,0.09530933035744561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,10240,0.4012471040089925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,8192,0.07512178023656209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,7168,0.15272177590264216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,7168,0.08187199963463677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,8192,0.3466977808210585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,6144,0.13110844294230142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,7168,0.07920977804395887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,6144,0.07221688826878865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,7168,0.3679182264539931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,6144,0.06975644164615206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,5120,0.11061955822838677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,6144,0.31831465827094185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,5120,0.06305511130226983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,4096,0.09429066711001927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,5120,0.05002133382691277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,4096,0.05889066722657946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,5120,0.35466665691799587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,3584,0.08208444383409289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,4096,0.040672000911500715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,4096,0.32293955485026044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,3584,0.047916445467207164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,3072,0.06865244441562228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,3584,0.043473776843812734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,3584,0.3362693256802029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,3072,0.043347554074393384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,2560,0.05850578016704983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,3072,0.032744887802335955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,3072,0.32048622767130536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,2560,0.03869955407248603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,2048,0.04796355631616381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,2560,0.03272711237271627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,2048,0.03437066740459866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,2560,0.3364933331807454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,1536,0.0373599992858039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,2048,0.02347999976740943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,2048,0.31682666142781574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,1536,0.02951644526587592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,1024,0.026712889472643535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,1536,0.01901155544651879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,1536,0.34063998858133954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,1024,0.023203555080625746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,768,0.020949333906173706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,1024,0.015265777707099915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,1024,0.2846631209055583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,768,0.01997511088848114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,512,0.014517333772447375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,768,0.013439999686347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,768,0.30693067444695366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,512,0.017310221989949543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,256,0.009125333693292405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,512,0.011462222370836469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,512,0.3100950982835558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,128,0.007621333003044128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,256,0.015200000670221118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,256,0.010121777653694153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,256,0.3307235505845812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,128,0.0147724449634552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,64,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,65536,128,0.01036888857682546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,65536,128,0.30465155177646214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,64,0.016214221715927124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,65536,32,0.018957333432303537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,65536,0.3226977719200982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,65536,0.1863457759221395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,16384,0.07962133487065633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,16384,0.057393776045905225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,65536,0.17431110805935332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,65536,32,0.0069164443347189165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,12288,0.0607831080754598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,16384,0.046033776468700834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,12288,0.04621600111325582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,16384,0.3320399920145671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,10240,0.05208977725770739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,12288,0.03491555651028951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,65536,0.33344443639119464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,12288,0.3381315602196588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,8192,0.042040000359217324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,10240,0.039579557047949895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,10240,0.3420257833268907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,8192,0.02625600000222524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,7168,0.03733777668741014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,8192,0.3192959891425239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,7168,0.0321413344807095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,10240,0.040780444939931236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,6144,0.033215999603271484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,7168,0.30979733996921116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,7168,0.022808889547983806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,6144,0.029810667037963867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,8192,0.035989334185918175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,5120,0.028174221515655518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,6144,0.32219910621643066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,5120,0.027414222558339436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,5120,0.3210231198204888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,5120,0.01797066628932953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,4096,0.022980444961124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,4096,0.026177777184380427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,4096,0.3252497778998481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,4096,0.015493333339691162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,3584,0.020987555384635925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,6144,0.02029511166943444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,3584,0.02167644434505039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,3072,0.018533334136009216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,3584,0.3250853220621745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,3584,0.016415999995337594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,3072,0.020317332612143624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,2560,0.016214221715927124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,3072,0.01313511116637124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,3072,0.3166293303171794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,2560,0.018918222851223417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,2048,0.014889778362380134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,2560,0.012449777788586087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,2560,0.33112088839213055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,2048,0.018204443984561496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,1536,0.010797333386209277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,2048,0.010692444112565784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,2048,0.3264053397708469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,1536,0.01680177781316969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,1024,0.008357333640257517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,1536,0.009562667045328353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,1536,0.33087288008795845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,1024,0.014090667168299357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,768,0.007052444749408298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,1024,0.3259066740671794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,768,0.01315111087428199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,768,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,512,0.004936000125275718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,512,0.012755555411179861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,512,0.33475377824571395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,512,0.006993778049945831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,256,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,768,0.32676532533433705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,256,0.012164444559150271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,256,0.00665244460105896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,128,0.0032435556252797446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,256,0.29120532671610516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,128,0.012119111087587146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,16384,128,0.28158667352464467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,64,0.003152888889114062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,128,0.00702044450574451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,16384,32,0.003261333331465721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,64,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,16384,32,0.013442666994200813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,65536,0.24790578418307832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,65536,0.1669928895102607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,16384,1024,0.007979555262459649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,65536,0.15148444970448813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,16384,0.05151199963357714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,65536,0.3394560019175212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,16384,0.33544267548455137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,12288,0.05101866523424784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,16384,0.0404151115152571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,12288,0.042131556404961475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,16384,0.06779822376039293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,10240,0.04394577940305074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,12288,0.031010667483011883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,12288,0.31111733118693036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,10240,0.03726666503482395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,8192,0.036455111371146307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,10240,0.027161777019500732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,8192,0.03325066632694668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,10240,0.3265191184149848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,7168,0.032073779238594904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,8192,0.022499556342760723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,8192,0.3141146765814887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,6144,0.0277804434299469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,7168,0.029917332861158583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,7168,0.020163555939992268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,7168,0.3138933446672228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,6144,0.026175111532211304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,6144,0.0203137778573566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,6144,0.3344257672627767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,5120,0.02383199996418423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,4096,0.020436444216304354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,5120,0.016364443633291457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,4096,0.02269866731431749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,5120,0.3791111045413547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,3584,0.017877333694034152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,4096,0.014499555031458536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,4096,0.3209795686933729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,3584,0.020414221617910597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,3072,0.015497777197096082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,5120,0.024135111106766596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,3584,0.33469687567816836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,3072,0.019186667270130582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,2560,0.01406133340464698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,3072,0.012449777788586087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,3072,0.31797954771253795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,2560,0.018601778480741713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,2560,0.3255315621693929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,2560,0.011745777395036487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,2048,0.013254221942689685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,2048,0.016915554801623028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,3584,0.014476444986131458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,2048,0.3258497714996338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,1536,0.01000088867213991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,2048,0.009333333207501305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,1536,0.015803555647532146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,1536,0.008344888687133789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,1024,0.00812266684240765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,1536,0.3321839968363444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,1024,0.013418667018413544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,768,0.006802666518423293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,1024,0.007576888634098901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,1024,0.3281955454084608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,768,0.012845333251688214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,512,0.004540444248252445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,768,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,768,0.3298177719116211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,512,0.012272888587580787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,256,0.003930666794379552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,512,0.006638222270541721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,512,0.3346159987979465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,256,0.012089778151777057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,128,0.0038533334930737815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,256,0.006342222293217977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,256,0.3344328933291965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,128,0.011756444142924415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,64,0.003216000066863166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,12288,128,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,12288,128,0.3191111087799072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,12288,32,0.003546666767862108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,64,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,12288,32,0.012437333663304647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,65536,0.20713422033521864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,65536,0.16030311584472656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,16384,0.05130399929152595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,65536,0.14225244522094727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,16384,0.04976088802019755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,16384,0.03870577613512675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,12288,0.040199110905329384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,16384,0.3367351161109076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,65536,0.34305156601799863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,12288,0.04078222314516703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,10240,0.03444177905718485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,12288,0.030919111437267725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,12288,0.3259093231625027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,10240,0.03592977921168009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,8192,0.028655999236636694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,10240,0.027119111683633592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,10240,0.3267555501725939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,8192,0.031983110639784075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,7168,0.025684444440735713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,8192,0.3255182107289632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,7168,0.02796088986926609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,7168,0.3292462295956082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,7168,0.019502222537994385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,6144,0.022665777140193518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,6144,0.02518400053183238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,6144,0.3249066670735677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,6144,0.01869244376818339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,5120,0.019488889310095046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,8192,0.024932444095611572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,5120,0.023271110322740342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,4096,0.01704355577627818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,5120,0.32130310270521373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,5120,0.01608000033431583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,4096,0.021967111362351313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,4096,0.014084445105658637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,3584,0.01482222146458096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,4096,0.3210844463772244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,3584,0.01963911122745938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,3072,0.012797333300113678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,3584,0.3459324571821425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,3584,0.013737777868906656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,3072,0.018727999594476488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,2560,0.011396444506115384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,3072,0.011401777466138205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,3072,0.33031998740302193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,2560,0.017786666750907898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,2048,0.011027555498811932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,2560,0.010680889089902243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,2560,0.3400088946024577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,2048,0.01683200067943997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,1536,0.008196444147162968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,2048,0.008978666530715095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,2048,0.32714666260613334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,1024,0.006672888994216919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,1536,0.014550222290886773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,1536,0.008022222254011366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,1536,0.3302328851487902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,1024,0.013401777380042605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,768,0.004911111046870549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,1024,0.007279111279381647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,1024,0.32119999991522896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,768,0.012758221891191272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,768,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,768,0.3276168770260281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,512,0.012138666378127204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,512,0.006640000061856375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,512,0.3177271154191759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,256,0.0038604442444112566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,256,0.012078222301271228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,256,0.34568354818556046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,256,0.006249777972698212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,128,0.003604444364706675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,512,0.0042160000238153665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,128,0.012135999898115793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,64,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,10240,128,0.310245328479343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,10240,128,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,64,0.012461333639091916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,10240,32,0.012075555821259817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,65536,0.16843288474612764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,65536,0.11083377732170953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,65536,0.07748533619774713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,16384,0.04666755596796671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,10240,32,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,65536,0.3134293291303846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,16384,0.037124445041020714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,12288,0.035926222801208496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,16384,0.3182693322499593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,16384,0.025390222668647766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,12288,0.03175910976197984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,10240,0.03073955575625102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,12288,0.018486221631368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,12288,0.32126400205824107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,10240,0.02777066661251916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,8192,0.025065778030289546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,10240,0.3274995485941569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,10240,0.01665599975321028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,8192,0.024496000674035814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,7168,0.022720000810093347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,8192,0.015271110667122735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,8192,0.31603556209140354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,7168,0.022022222479184467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,6144,0.0206631107462777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,7168,0.013422222601042854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,7168,0.3194719950358073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,6144,0.02034222251839108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,6144,0.01260355528857973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,5120,0.017228444417317707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,6144,0.3227546744876438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,5120,0.019357333580652874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,4096,0.015200000670221118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,5120,0.011420444481902652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,5120,0.33143732282850474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,4096,0.019010666343900893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,4096,0.010010666317409938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,4096,0.3308648798200819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,3584,0.016892444756295946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,3584,0.010050666828950247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,3584,0.3215831120808919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,3072,0.011809777882364062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,3072,0.016201777590645682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,3072,0.33056889639960396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,3072,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,2560,0.010012444522645738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,2560,0.016200888488027785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,2560,0.008592888712882996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,2560,0.323524448606703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,2048,0.009959999885823991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,2048,0.014490667316648694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,2048,0.3184000121222602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,2048,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,1536,0.007271111011505127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,1536,0.013510222236315409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,1536,0.31755733489990234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,1024,0.005680888891220093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,3584,0.013140444126394061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,1024,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,1024,0.012815111213260226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,768,0.004589333302444882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,1024,0.31426933076646596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,768,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,768,0.012524444195959302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,512,0.004090666770935059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,768,0.321723567114936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,512,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,512,0.0058986664646201665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,256,0.003646222253640493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,512,0.3196764522128635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,256,0.01187999960449007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,128,0.003212444484233856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,256,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,256,0.3892666763729519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,128,0.011695110963450538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,64,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,8192,128,0.29450755649142796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,128,0.005994666781690385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,8192,32,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,64,0.011750222080283694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,65536,0.14322043789757624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,8192,32,0.012137778103351593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,65536,0.10261244244045681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,16384,0.04446666770511203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,65536,0.07576888799667358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,8192,1536,0.0069715554515520734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,16384,0.035360889302359685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,16384,0.023443554838498432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,12288,0.035643554396099515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,16384,0.3595164351993137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,65536,0.3366897900899251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,12288,0.030157334274715845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,12288,0.018037334084510803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,10240,0.031096001466115315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,12288,0.34142133924696183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,10240,0.02500266664557987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,10240,0.016207110550668504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,10240,0.3508124351501465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,8192,0.022857778602176245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,8192,0.014807111687130399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,7168,0.021331555313534204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,7168,0.02125955621401469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,8192,0.023211555348502264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,7168,0.34663645426432294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,6144,0.018975110517607797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,7168,0.013069333301650153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,6144,0.019967110620604623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,8192,0.3292026784684923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,6144,0.32161688804626465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,5120,0.016203555795881484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,6144,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,5120,0.01868088874551985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,5120,0.011119999819331698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,4096,0.014808000789748298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,5120,0.32455645667182076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,4096,0.018741332822375827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,3584,0.01238044434123569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,4096,0.009877333210574256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,4096,0.3300444549984402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,3584,0.01703377730316586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,3072,0.011202666494581433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,3584,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,3584,0.32846400472852916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,3072,0.01624000072479248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,2560,0.010066666536860997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,3072,0.008012444608741337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,3072,0.3195902241600884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,2560,0.01532444523440467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,2560,0.007655111451943715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,2048,0.009394666386975182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,2560,0.3309253321753608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,2048,0.014383110735151501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,1536,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,2048,0.007327111230956183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,2048,0.3415084415011936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,1536,0.013447999954223633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,1024,0.0052471111218134565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,1536,0.32807734277513295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,1024,0.012652444342772165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,1024,0.006251555350091722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,1024,0.3200959894392225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,768,0.004260444392760594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,768,0.01218311074707243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,768,0.3349946604834662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,768,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,1536,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,512,0.011755555868148804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,512,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,256,0.003638222399685118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,512,0.33615732192993164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,256,0.011768889096048145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,256,0.005646222167544895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,128,0.0032071111102898917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,512,0.003887999802827835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,128,0.011432888607184092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,7168,128,0.005899555567238066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,128,0.30639200740390354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,64,0.002946666752298673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,7168,32,0.0032266666077905228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,64,0.012136889000733694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,7168,32,0.012468444804350534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,65536,0.11849510669708252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,7168,256,0.3262106577555339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,65536,0.09548978010813396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,16384,0.03624533282385932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,65536,0.07296622461742826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,16384,0.03317600157525804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,12288,0.029859556092156306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,16384,0.32500267028808594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,65536,0.33675821622212726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,12288,0.02753866712252299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,10240,0.025795555777019922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,12288,0.017363554901546903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,12288,0.3334355619218614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,10240,0.024806222981876794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,8192,0.02203111184967889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,10240,0.015837333268589444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,10240,0.3324844572279188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,8192,0.014175999495718213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,8192,0.33708445231119794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,7168,0.02017422186003791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,16384,0.02148888839615716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,7168,0.020169778002632987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,6144,0.017954667409261067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,7168,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,7168,0.32779733339945477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,6144,0.01902577777703603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,6144,0.011753777662913004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,5120,0.015522667103343539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,6144,0.39691821734110516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,8192,0.022167111436525982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,5120,0.018216000662909616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,4096,0.014448000325096978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,5120,0.01036533299419615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,5120,0.27544622951083714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,4096,0.01792266633775499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,4096,0.008985777695973715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,3584,0.012059555285506778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,4096,0.37032800250583225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,3584,0.01620444489849938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,3072,0.01071200023094813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,3584,0.00867733359336853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,3584,0.34153244230482316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,3072,0.015154666370815702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,2560,0.009496889180607265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,3072,0.3387377791934543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,2560,0.01426755554146237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,2560,0.33719645606146914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,2560,0.00738933351304796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,2048,0.00815466625822915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,2048,0.01313333296113544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,2048,0.341371562745836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,2048,0.0069875551594628235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,1536,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,3072,0.007837333612971837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,1536,0.012823111481136747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,1024,0.0046880000995265115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,1536,0.33214134640163845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,1024,0.012418666647540199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,1024,0.3312151167127821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,1024,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,768,0.004222222086456087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,768,0.012082666986518435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,768,0.33064177301194936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,768,0.005969777703285217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,512,0.003912000192536248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,1536,0.006568000134494569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,512,0.33415643374125165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,512,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,256,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,256,0.011741333537631564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,256,0.33267733785841197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,256,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,128,0.0031537777847713898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,128,0.011762667033407422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,6144,128,0.3051973448859321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,64,0.0028826666788922418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,6144,128,0.00563733321097162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,512,0.011782222323947482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,6144,32,0.0031582222630580268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,65536,0.1029768917295668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,32,0.011678222152921887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,65536,0.08867022063997056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,65536,0.070269332991706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,16384,0.03266755408710904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,65536,0.33666398790147567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,16384,0.03156266609827677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,6144,64,0.011554666691356234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,16384,0.32500799496968585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,12288,0.02535111043188307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,16384,0.020972443951500788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,12288,0.025949334104855854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,12288,0.016873778568373788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,10240,0.022408000297016565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,12288,0.34090222252739805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,10240,0.02253955602645874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,8192,0.018967999352349173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,10240,0.015474667151769003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,10240,0.32819732030232746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,8192,0.021412443783548143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,7168,0.0178195552693473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,8192,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,8192,0.3431973457336426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,7168,0.0192275560564465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,6144,0.01642933322323693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,7168,0.012456888953844706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,7168,0.3292488786909315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,6144,0.01851288808716668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,5120,0.01370844410525428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,6144,0.32515645027160645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,6144,0.01148799972401725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,5120,0.01734755602147844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,4096,0.012718222207493253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,5120,0.010120000276300643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,5120,0.3136986626519097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,4096,0.017445334129863314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,3584,0.01071111112833023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,4096,0.008593777815500895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,3584,0.015885333220163982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,3072,0.009655111365848118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,3584,0.00796088907453749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,3584,0.3122444417741564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,3072,0.014377777775128683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,2560,0.008952000074916417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,3072,0.007640000018808577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,3072,0.3338649008009169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,2560,0.01348355578051673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,2560,0.007287999822033777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,2560,0.32600177658928764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,2048,0.00868000007337994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,2048,0.013205333716339536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,2048,0.34223466449313694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,2048,0.006994666324721442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,1536,0.0052977779673205475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,1536,0.01276000009642707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,1536,0.33565600713094074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,1536,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,1024,0.004613333278232151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,1024,0.012418666647540199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,1024,0.3410800033145481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,1024,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,4096,0.4143271181318495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,768,0.012080888781282636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,768,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,512,0.003909333298603694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,768,0.3866133424970839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,512,0.011776888536082374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,256,0.00319377767542998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,512,0.3681617842780219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,768,0.003903111235962974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,256,0.011699555648697747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,128,0.003177777760558658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,256,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,256,0.3519502215915256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,128,0.011020444333553314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,64,0.002883555574549569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,5120,128,0.005570666657553778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,5120,128,0.3153982162475586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,64,0.011063999599880643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,5120,32,0.01184622198343277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,65536,0.08478666676415338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,65536,0.08418666654162937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,65536,0.06774755318959554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,65536,0.3327084382375081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,16384,0.027075555589463975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,5120,32,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,16384,0.029654221402274236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,12288,0.023071999351183575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,16384,0.3350497881571452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,16384,0.02030044462945726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,12288,0.023620444867346022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,12288,0.016543999314308167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,10240,0.01960444450378418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,12288,0.32767823007371694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,10240,0.02163644466135237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,8192,0.017852443787786696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,10240,0.014911111858155994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,10240,0.33344088660346133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,8192,0.020421332783169217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,7168,0.015417777829700045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,8192,0.01310755560795466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,8192,0.3390000131395128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,7168,0.01867555578549703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,7168,0.011415999796655444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,7168,0.33420266045464414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,6144,0.01788444485929277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,6144,0.010440000229411654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,6144,0.31754933463202584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,5120,0.016726222303178575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,6144,0.013526221944226159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,5120,0.009152000149091085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,4096,0.010272000398900773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,4096,0.016187555260128446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,5120,0.011046222514576383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,4096,0.008351999852392409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,3584,0.008805333740181392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,4096,0.32886044184366864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,3584,0.014857777290874057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,3072,0.008186666501892937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,5120,0.32366400294833714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,3072,0.013768889009952545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,3584,0.4713039928012424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,2560,0.007280888656775157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,3072,0.3225991196102566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,3072,0.007696888513035244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,2560,0.013546666337384118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,2048,0.006330666856633292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,2560,0.3095040056440565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,2048,0.013163555827405719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,2048,0.3079084555308024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,3584,0.00796977761718962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,1536,0.005225777626037598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,2048,0.006960888703664144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,1536,0.01277599980433782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,1536,0.006631999793979857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,1024,0.004222222086456087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,1536,0.3144773377312554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,1024,0.012467555701732635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,768,0.003839111162556542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,1024,0.006247111078765657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,2560,0.007289778027269576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,1024,0.330260435740153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,512,0.0038204445607132385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,768,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,768,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,512,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,768,0.32693423165215385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,256,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,512,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,512,0.3545573287540012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,256,0.011375111010339526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,128,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,256,0.005581333405441708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,256,0.331059561835395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,128,0.011054221954610614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,64,0.0031635556370019913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,4096,128,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,4096,128,0.3220746782090929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,4096,32,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,64,0.011061333119869232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,65536,0.07380177577336629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,4096,32,0.012062222593360476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,65536,0.0821288890308804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,16384,0.023732443650563557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,65536,0.06655110915501912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,65536,0.34354755613538956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,16384,0.027726221415731642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,12288,0.019652444455358718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,16384,0.33973333570692277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,16384,0.019991111424234178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,12288,0.023367111881573994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,10240,0.016968000266287062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,12288,0.016494222813182406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,12288,0.33605511983235675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,10240,0.020982222424613103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,8192,0.01516266663869222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,10240,0.014847111370828418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,10240,0.333771546681722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,8192,0.019638222124841478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,8192,0.012509333590666452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,7168,0.014509333504570855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,8192,0.34116265508863663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,7168,0.018713777263959248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,6144,0.01290933373901579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,7168,0.011448000040319232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,7168,0.3368328942192926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,6144,0.01752800080511305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,5120,0.010407999985747868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,6144,0.010129777921570672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,6144,0.36805423100789386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,5120,0.016607999801635742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,4096,0.009692444569534725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,5120,0.009316444396972656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,5120,0.3338453239864773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,4096,0.014800000521871777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,3584,0.00867733359336853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,4096,0.00830666638082928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,4096,0.34450933668348527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,3584,0.014119111829333834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,3072,0.008063111040327284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,3584,0.008023111356629265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,3584,0.31073421902126735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,3072,0.013780444032616086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,2560,0.00703111125363244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,3072,0.007601777712504069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,3072,0.33068622483147514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,2560,0.31500177913241917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,2048,0.005910222315125995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,2048,0.012786666552225748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,2048,0.3211351235707601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,2048,0.006989333364698622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,2560,0.013548444542619916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,1536,0.004543111142185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,2560,0.007337777978844113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,1536,0.01274577776590983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,1024,0.004206222378545337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,1536,0.31793689727783203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,1536,0.006643555644485686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,1024,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,768,0.0038613333470291565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,1024,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,1024,0.32301333215501576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,768,0.012122666670216454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,512,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,768,0.005944889038801193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,768,0.34008532100253636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,512,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,256,0.003519999898142285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,512,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,512,0.323745780520969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,128,0.0029004443850782183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,256,0.005799110978841782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,256,0.34808532396952313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,128,0.011047999891969891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,64,0.0028693332440323303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3584,128,0.005640000104904175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3584,32,0.0029004443850782183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,64,0.011075555450386472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,65536,0.0641546646753947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,32,0.01182933317290412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,65536,0.08004266685909696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3584,256,0.011742221812407175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,16384,0.02139733400609758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,65536,0.0654568870862325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,65536,0.3343324396345351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,16384,0.026593777868482802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,16384,0.342196438047621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,12288,0.01925244430700938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3584,128,0.41632265514797634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,12288,0.022653333014912073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,12288,0.016183111402723525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,12288,0.4037200080023871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,10240,0.02089333368672265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,16384,0.019661333825853135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,10240,0.39719022644890684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,10240,0.014173333843549093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,8192,0.013172444370057849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,8192,0.019648000597953796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,10240,0.015575110912322998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,8192,0.36454399426778156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,8192,0.011795555551846823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,7168,0.012478222449620565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,7168,0.018046221799320646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,7168,0.010743111371994019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,6144,0.012180444267061023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,7168,0.33821866247389054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,6144,0.01686755485004849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,5120,0.010242666635248397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,6144,0.010020444790522257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,6144,0.3620764414469401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,5120,0.015863110621770222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,4096,0.009327111144860586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,5120,0.32573334376017254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,4096,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,4096,0.324614233440823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,4096,0.008229333493444655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,3584,0.008047999607192146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,3584,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,3584,0.32044800122578937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,3584,0.008020444048775567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,3072,0.007613333563009898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,3072,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,5120,0.009343110852771336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,2560,0.007296000089910295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,3072,0.31752443313598633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,3072,0.007630222373538547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,2560,0.013175110850069257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,2048,0.006263111200597551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,2560,0.3173457781473796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,2048,0.013019555144839816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,2048,0.31594665845235187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,1536,0.005241777747869492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,1536,0.012747555143303342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,1536,0.3279422124226888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,2560,0.007308444215191736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,1024,0.004209777961174647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,1536,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,1024,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,2048,0.007019555403126611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,1024,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,1024,0.3306097719404432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,768,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,768,0.33204711808098686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,768,0.005897777775923411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,512,0.0035235554807715942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,512,0.011727111207114326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,512,0.32494134373135036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,512,0.0058062221441004015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,256,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,256,0.011425777441925473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,768,0.003935111065705617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,128,0.0031866667171319327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,256,0.31749600834316677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,256,0.005616000129116907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,128,0.011758222348160215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,3072,128,0.005590222362014983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,64,0.0030337776988744736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,3072,32,0.0029360000044107437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,3072,128,0.28705332014295787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,65536,0.05918311410480075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,32,0.011439111497667102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,65536,0.0773635572857327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,16384,0.019824888971116807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,65536,0.06404711140526666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,16384,0.026181333594852026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,65536,0.468065791659885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,16384,0.019503111640612285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,16384,0.33487733205159503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,12288,0.01731911136044396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,12288,0.022706665926509436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,10240,0.014497778481907315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,12288,0.01571999986966451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,12288,0.33675644132826066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,3072,64,0.011056000159846412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,10240,0.020675554871559143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,8192,0.01219733307758967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,10240,0.01311644415060679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,8192,0.01924888955222236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,10240,0.3859848976135254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,8192,0.011111111276679568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,8192,0.33928444650438094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,7168,0.018031110366185505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,6144,0.011414222419261932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,7168,0.010431999961535135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,7168,0.33650578392876523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,6144,0.016470222009552848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,6144,0.3405297862158881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,6144,0.009776888622177972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,5120,0.01514933341079288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,7168,0.011528888510333167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,5120,0.008980444735950893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,4096,0.009094222552246517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,5120,0.3448586728837755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,4096,0.014454222387737699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,4096,0.3447226683298747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,4096,0.00832799987660514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,3584,0.007975999679830339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,5120,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,3584,0.014138667119873894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,3584,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,3072,0.007329777710967594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,3584,0.3434782293107774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,3072,0.013507555756303998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,2560,0.006836444553401735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,3072,0.007711110843552484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,3072,0.3303644392225477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,2560,0.013166222307417126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,2048,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,2560,0.34332532352871364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,2048,0.012809777425395118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,2048,0.00700711127784517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,2048,0.33187466197543675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,1536,0.004905777672926585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,1536,0.012480888929631976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,1536,0.34175732400682235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,1536,0.00666933341158761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,1024,0.0042453333735466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,2560,0.007283555964628856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,1024,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,768,0.0038657776183552216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,1024,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,1024,0.3249439928266737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,768,0.012111999922328524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,512,0.003531555748648114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,768,0.33364176750183105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,512,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,512,0.005946666830115848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,256,0.0032471112079090546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,256,0.011725333001878528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,256,0.3353706730736627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,768,0.005922666854328579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,128,0.003173333282272021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,256,0.005620444400442972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,128,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,512,0.31551289558410645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,64,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,64,0.011420444481902652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2560,32,0.00313688897424274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2560,32,0.012079999678664737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,65536,0.04634666774008009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,65536,0.07643289036220975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2560,128,0.005655999812814925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2560,128,0.3126800060272217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,16384,0.018283555905024212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,65536,0.06297244628270467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,65536,0.33465509944491917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,16384,0.025807110799683466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,12288,0.012898666991127862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,16384,0.018946665856573317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,16384,0.32451732953389484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,12288,0.014138667119873894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,12288,0.022289777795473736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,12288,0.33294489648607045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,10240,0.020262221495310467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,10240,0.012787555654843649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,10240,0.4715004497104221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,8192,0.3847831090291341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,10240,0.013592888911565145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,7168,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,7168,0.01683200067943997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,8192,0.012878222597969903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,7168,0.010423111418883005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,8192,0.018538667096032035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,6144,0.01070844464831882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,7168,0.30965865982903373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,6144,0.015896888242827523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,5120,0.0099973330895106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,6144,0.32323288917541504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,5120,0.014928888943460254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,8192,0.011104889214038849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,4096,0.008663111262851292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,5120,0.00920800036854214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,5120,0.3203128973642985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,4096,0.008323555191357931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,3584,0.00738666703303655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,4096,0.33637245496114093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,3584,0.014158222410413953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,3584,0.008086221913496653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,6144,0.009780444204807281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,3584,0.32948443624708385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,3072,0.013427555561065674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,3072,0.32345954577128094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,3072,0.007663110891977946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,2560,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,4096,0.014164444473054675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,2560,0.3262737856970893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,2560,0.00742133292886946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,2048,0.0053004444473319584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,3072,0.006631999793979857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,2048,0.012809777425395118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,2048,0.30725333425733775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,2048,0.00702311098575592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,1536,0.004618666652176115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,1536,0.012448000411192576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,2560,0.013097777962684631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,1024,0.004210666649871402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,1536,0.3269760078854031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,1536,0.006937777830494775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,1024,0.012110222544935016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,768,0.003910222401221593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,1024,0.006232000059551663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,1024,0.32144443194071454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,768,0.011869333684444427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,512,0.0035146665241983202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,768,0.005962666538026597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,768,0.33664533827039933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,512,0.011818666425016193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,256,0.0031866667171319327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,512,0.005604444278611078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,512,0.32946933640374076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,256,0.011394666300879585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,128,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,256,0.005558222118351195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,256,0.3300204541948107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,128,0.011076444553004371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,64,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,2048,128,0.005568888866239124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,2048,32,0.002936888900068071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,2048,128,0.30621422661675346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,65536,0.036666668123669095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,32,0.011777777638700275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,65536,0.07457688781950209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,65536,0.06014577547709147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,65536,0.3364444573720296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,16384,0.01425066590309143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,16384,0.02530577778816223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,16384,0.3340880076090495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,16384,0.017601778109868366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,2048,64,0.01107022249036365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,12288,0.021693333983421326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,12288,0.013812444276279874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,12288,0.33821866247389054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,10240,0.011439111497667102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,10240,0.019474666979577806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,10240,0.33971820937262637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,8192,0.01054044481780794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,12288,0.01293688929743237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,8192,0.017095110482639737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,8192,0.011125333607196808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,7168,0.010096000300513374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,7168,0.01620533400111728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,10240,0.012462221913867526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,7168,0.3398755656348334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,6144,0.010060444474220276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,7168,0.010403555300500657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,6144,0.01585155559910668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,8192,0.32611200544569224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,6144,0.3356248802608914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,6144,0.009696889254781935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,5120,0.009475555684831407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,5120,0.01482222146458096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,5120,0.009017777939637503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,4096,0.009223110973834991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,5120,0.32849333021375865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,3584,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,4096,0.014138667119873894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,4096,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,3584,0.013778666655222574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,3584,0.008086221913496653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,4096,0.46393691168891055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,3584,0.339488877190484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,3072,0.01346666696998808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,2560,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,3072,0.007603555917739868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,2560,0.013080889152155982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,3072,0.42106045616997617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,2048,0.0052826665341854095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,2560,0.2984542316860623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,2560,0.007297777467303806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,2048,0.006913777854707506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,2048,0.012839111189047495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,1536,0.0048142220411035745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,2048,0.30273956722683376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,3072,0.006632888896597757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,1536,0.012850667039553324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,1024,0.004230222354332606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,1536,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,1536,0.31357423464457196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,768,0.0038604442444112566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,1024,0.012163555456532372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,1024,0.006100444330109491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,1024,0.3183022340138753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,768,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,768,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,768,0.36494133207533097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,512,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,512,0.3297457695007324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,512,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,256,0.0032106666929192017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,256,0.011391110718250275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,256,0.329128000471327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,256,0.005587555468082428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,128,0.0028497777465316984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,128,0.01108977778090371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,512,0.0035253332720862497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,64,0.00281866660548581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1536,128,0.2984657817416721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1536,32,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,64,0.011194667054547204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1536,32,0.011797332929240333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,65536,0.02602133320437537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,65536,0.07165600193871392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,65536,0.31157512134975857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,16384,0.012610666453838348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,65536,0.05786399708853828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,16384,0.02464977734618717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1536,128,0.00562755556570159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,16384,0.016545777519543965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,12288,0.019914666811625164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,12288,0.3348737822638617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,12288,0.013833777772055732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,10240,0.010393777655230628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,16384,0.33020000987582737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,12288,0.012055111428101858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,10240,0.012402666939629449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,8192,0.00980533328321245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,8192,0.01664177742269304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,8192,0.3449110984802246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,10240,0.018648000227080453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,8192,0.011071110765139261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,7168,0.009527111219035255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,7168,0.016262221667501662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,10240,0.3247048854827881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,6144,0.00830666638082928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,7168,0.33251823319329155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,7168,0.010375111467308467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,6144,0.01552977826860216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,6144,0.009686222506894005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,5120,0.0087333329849773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,6144,0.32140888108147514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,5120,0.014831999937693277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,4096,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,5120,0.009085333181752099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,5120,0.3285333315531413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,4096,0.014127999544143677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,3584,0.006949333681000604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,4096,0.32280977567036945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,3584,0.013781333135233985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,3584,0.008013333711359236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,3584,0.3084302213456896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,3072,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,3072,0.013468444347381592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,3072,0.336408880021837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,3072,0.007620444728268518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,2560,0.005585777676767773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,4096,0.008290666672918532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,2560,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,2048,0.005220444252093633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,2560,0.3226355446709527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,2048,0.01275466630856196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,2560,0.007423111134105259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,1536,0.004763555609517627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,2048,0.006969778074158563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,2048,0.30024266242980957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,1024,0.0038568890757030914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,1536,0.012840000291665396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,1536,0.006604444649484422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,1536,0.32819822099473744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,1024,0.012411555482281579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,768,0.003528888854715559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,1024,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,768,0.012051555845472546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,512,0.0035537779331207275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,768,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,1024,0.42737423049079043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,768,0.323341343137953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,256,0.0031999999450312722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,512,0.011722666521867117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,512,0.0059368887709246735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,256,0.01179377817445331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,128,0.0028311111446883944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,256,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,256,0.29900532298617893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,128,0.011504888534545898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,64,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,1024,128,0.005559111220969095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,1024,32,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,128,0.34112177954779727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,64,0.011372444530328115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,65536,0.023222222924232483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,1024,32,0.01184177812602785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,65536,0.0687075588438246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,65536,0.05756800042258369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,16384,0.011536888778209686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,65536,0.355605337354872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,16384,0.024089778463045757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,1024,512,0.3289946715037028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,12288,0.01000355515215132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,16384,0.3358026610480414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,16384,0.016520000166363187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,12288,0.01974311139848497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,12288,0.013778666655222574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,10240,0.010370666782061258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,12288,0.3288808928595649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,10240,0.018518222702874076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,8192,0.008753777378135258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,10240,0.012438221938080259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,10240,0.32712888717651367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,8192,0.016890666551060148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,7168,0.00832088871134652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,8192,0.3411173290676541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,7168,0.01627288924323188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,7168,0.010476444330480365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,6144,0.008797333472304875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,6144,0.015519110692871941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,6144,0.3159315586090088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,8192,0.011356444822417365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,5120,0.007961778177155389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,6144,0.009673777553770278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,5120,0.014856000741322836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,7168,0.3286657863193088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,5120,0.3299200005001492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,4096,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,4096,0.014182221558358936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,4096,0.32472533649868435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,3584,0.007010666860474481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,3584,0.013808888693650564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,3584,0.31273243162367076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,5120,0.009190222455395592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,3072,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,3584,0.00795911086930169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,3072,0.01344622257683012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,4096,0.007998222278224098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,3072,0.32381243175930446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,3072,0.007624000310897827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,2560,0.00591822216908137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,2560,0.013111111190583972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,2560,0.007255111303594377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,2048,0.005207111024194294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,2560,0.3233608934614393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,2048,0.013071111506885953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,1536,0.004538666870858935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,2048,0.007000000112586551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,2048,0.3306809001498752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,1536,0.012730666332774691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,1024,0.003863111138343811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,1536,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,1536,0.321636438369751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,1024,0.01238222254647149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,768,0.003583111282851961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,1024,0.006227555374304454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,1024,0.32573066817389595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,768,0.01204533295498954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,512,0.003294222263826264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,768,0.005896889087226655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,768,0.3170062171088325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,512,0.011762667033407422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,256,0.003195555466744635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,512,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,512,0.33145154847039116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,256,0.011775111158688864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,128,0.0029804443733559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,256,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,256,0.3160995642344157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,128,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,64,0.0029173334025674393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,768,128,0.005561777700980504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,768,32,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,768,128,0.30555823114183217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,64,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,768,32,0.011074666347768573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,65536,0.06740888622072008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,16384,0.010402667025725046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,65536,0.0556817783249749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,65536,0.33463290002610946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,16384,0.023021333747439917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,12288,0.010145777629481422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,16384,0.01626755628320906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,12288,0.0195777780479855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,16384,0.45511378182305234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,12288,0.013759111364682516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,12288,0.3217208915286594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,65536,0.017424000634087455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,10240,0.018600889378123812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,8192,0.008309333688682979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,10240,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,10240,0.32066933314005536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,8192,0.016857778032620747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,7168,0.007963555554548899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,8192,0.0107022225856781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,8192,0.3146266672346327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,7168,0.016223111086421542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,7168,0.010341333018408881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,7168,0.3202435440487332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,6144,0.015441776977645027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,10240,0.009123555488056606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,6144,0.32042045063442653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,5120,0.007983110845088959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,6144,0.009710222482681274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,5120,0.014530667000346713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,5120,0.009002666506502364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,5120,0.3826124403211806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,4096,0.007349333001507654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,6144,0.008962666822804345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,4096,0.014161777165200977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,3584,0.00702311098575592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,4096,0.3399466673533122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,3584,0.013811555173661975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,3584,0.32585954666137695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,3584,0.008036444584528605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,3072,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,3072,0.013467555244763693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,3072,0.3213626543680827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,3072,0.007629333270920648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,2560,0.005567111074924469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,4096,0.008035555481910706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,2560,0.01313866674900055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,2048,0.005223999834722943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,2560,0.31612356503804523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,2560,0.007296000089910295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,2048,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,2048,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,1536,0.004609777695602841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,2048,0.31387021806504994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,1536,0.012791111237472959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,1024,0.003966222206751506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,1536,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,1536,0.3566817707485623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,1024,0.012413333687517377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,768,0.0035413333939181436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,1024,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,1024,0.314759996202257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,768,0.012087110843923358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,512,0.003175111073586676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,768,0.005912000106440649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,768,0.3499893347422282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,512,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,256,0.003168888803985384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,512,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,512,0.3091368940141466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,256,0.01180622229973475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,128,0.00300799993177255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,256,0.005581333405441708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,256,0.34012889862060547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,128,0.011408000356621213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,64,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,512,128,0.3011404408348931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,512,32,0.0029013332807355454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,64,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,65536,0.014967999524540372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,512,32,0.011039999624093374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,65536,0.06607288784450956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,65536,0.3316195540957981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,16384,0.009681777821646797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,16384,0.022290666898091633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,512,128,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,16384,0.3291155497233073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,16384,0.01650311052799225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,12288,0.009351111120647853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,12288,0.019634667370054457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,12288,0.32492444250318736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,65536,0.04967022273275587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,10240,0.008671999805503422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,10240,0.018312000566058688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,10240,0.31457334094577366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,10240,0.012448888685968188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,8192,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,8192,0.01683999929163191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,8192,0.3207804361979167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,8192,0.011033777561452655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,7168,0.0076248885856734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,7168,0.016191111670600045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,12288,0.013596444494194455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,6144,0.007344889144102733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,7168,0.3163928985595703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,7168,0.010397333237859938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,6144,0.015538665983412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,6144,0.009705777797434065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,6144,0.31350933180914986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,5120,0.014944000376595391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,4096,0.0069360000391801195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,5120,0.009008000294367472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,4096,0.014461333552996317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,5120,0.47979646258884007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,3584,0.006581333362393909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,4096,0.3222568829854329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,4096,0.008022222254011366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,3584,0.013793778088357715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,3584,0.00794400026400884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,3072,0.006918222126033571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,3584,0.3206417825486925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,3072,0.013759111364682516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,5120,0.007290666302045186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,2560,0.006638222270541721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,3072,0.007307555940416124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,3072,0.32132355372111004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,2560,0.01313333296113544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,2048,0.006303999986913469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,2560,0.007252444823582967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,2560,0.32705333497789174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,2048,0.012784889174832238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,1536,0.004555555681387584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,2048,0.006735111276308696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,2048,0.32662667168511283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,1536,0.012463111016485425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,1024,0.003949333396222857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,1536,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,1536,0.324106666776869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,1024,0.012098666694429187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,1024,0.006239111224810283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,1024,0.33193956481085884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,768,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,768,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,512,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,768,0.33613332112630206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,512,0.011782222323947482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,512,0.3286995622846815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,512,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,768,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,256,0.011717333561844297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,256,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,128,0.002893333426780171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,256,0.34696443875630695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,128,0.011400889191362592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,256,128,0.32347022162543404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,256,128,0.005583111196756363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,64,0.0028719999310043124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,32,0.002855111120475663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,256,256,0.0032097777972618737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,64,0.011414222419261932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,65536,0.01481599940194024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,256,32,0.011422221859296164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,65536,0.06537866592407227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,16384,0.008982222113344405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,65536,0.33803643120659727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,16384,0.023079999619060095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,16384,0.32824622260199654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,12288,0.00793511089351442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,12288,0.019293333093325298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,12288,0.3334044350518121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,12288,0.01346577786737018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,10240,0.007377777662542131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,16384,0.016509332590632968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,10240,0.018022222651375663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,10240,0.012392889294359418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,8192,0.007284444239404466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,10240,0.3278506596883138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,8192,0.0166275550921758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,65536,0.04840355449252658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,7168,0.006663110935025745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,8192,0.010728889041476779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,7168,0.016148444679048326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,7168,0.010390222072601318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,7168,0.33663733800252277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,6144,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,6144,0.015478221906556023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,6144,0.3408835464053684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,5120,0.006309333360857434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,8192,0.3167031076219347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,5120,0.014811555544535318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,5120,0.00905777762333552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,5120,0.32134487893846303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,4096,0.014182221558358936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,6144,0.009685333404276106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,4096,0.27982576688130695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,3584,0.0063982220987478895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,4096,0.008043555749787224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,3584,0.014081777797804939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,3072,0.005974222388532426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,3584,0.3293235566880968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,3584,0.007727111379305522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,3072,0.013761777844693927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,4096,0.006582222051090664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,2560,0.006598222172922558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,3072,0.007409777906205919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,3072,0.2991360028584798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,2560,0.01316177762216992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,2048,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,2560,0.006992888947327931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,2560,0.2919813262091743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,1536,0.004545777622196409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,2048,0.013079110946920184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,2048,0.006667555620272954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,2048,0.31118665801154244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,1536,0.012748444245921241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,1024,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,1536,0.006325333482689328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,1536,0.3340755568610297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,1024,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,768,0.0035751110149754416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,1024,0.006130666782458623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,768,0.012106666962305704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,512,0.0031795555518733132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,768,0.005977777971161737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,768,0.32613245646158856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,512,0.011796444654464722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,512,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,256,0.003169777699642711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,512,0.35095821486579043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,256,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,256,0.33919111887613934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,256,0.005565333283609814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,128,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,1024,0.30988976690504283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,128,0.011563555234008364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,64,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,48,128,128,0.005548444473081165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,48,128,128,0.3152657879723443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,64,0.011398221883508893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,65536,0.012763555679056378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,128,32,0.01036888857682546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,16384,0.007697777615653143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,65536,0.06327911218007405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,12288,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,16384,0.02239111065864563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,10240,0.006604444649484422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,12288,0.01962844365172916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,8192,0.0069324444565508105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,10240,0.01791466606987847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,7168,0.006313777632183499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,8192,0.01688799924320645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,6144,0.006608888920810487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,128,32,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,7168,0.015843555331230164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,6144,0.015573332707087198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,4096,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,5120,0.014450665977266101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,3584,0.006291555447710886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,4096,0.01443199978934394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,3072,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,3584,0.013447999954223633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,2560,0.006285333385070165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,3072,0.01349511080318027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,2048,0.00600000015563435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,2560,0.012832000023788877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,1536,0.004559999952713649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,2048,0.013125333521101209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,1024,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,5120,0.005967111223273807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,1536,0.012142221960756512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,768,0.0035662220584021676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,512,0.003485333174467087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,1024,0.01239911135700014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,256,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,768,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,128,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,512,0.011733333269755045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,64,0.002879111096262932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,256,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,64,32,0.002508444504605399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,128,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,65536,0.011776000261306763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,64,0.011090666883521609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,16384,0.007958222594526079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,64,32,0.01180266671710544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,12288,0.007620444728268518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,65536,0.06022310919231839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,16384,0.022407111194398668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,12288,0.019716444942686293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,8192,0.006950221955776215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,10240,0.018040888839297824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,7168,0.006593777901596493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,8192,0.016866667403115165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,6144,0.006288888967699475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,7168,0.015864888827006023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,5120,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,6144,0.015487111277050443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,4096,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,5120,0.014837332897716098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,3584,0.005981333139869902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,4096,0.014456889695591398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,3072,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,3584,0.013920888304710388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,10240,0.006968888971540663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,2560,0.006278222219811545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,2048,0.00599644457300504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,3072,0.013757333159446716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,1536,0.004557333472702238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,2560,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,1024,0.003887999802827835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,2048,0.012778666284349231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,768,0.003593777616818746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,1536,0.012543111211723752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,512,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,1024,0.011815110842386881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,256,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,768,0.011934222446547614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,128,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,512,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,64,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,256,0.01143022212717268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,48,32,32,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,128,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,64,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,48,32,32,0.01034755590889189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,16384,0.350492450926039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,16384,0.1649466620551215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,16384,0.14749244848887125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,12288,0.2643164528740777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,12288,0.12958844502766928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,12288,0.1298417780134413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,16384,0.3547617859310574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,10240,0.21534400516086152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,10240,0.10909066597620647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,12288,0.5017120043436686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,10240,0.09290399816301133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,8192,0.09483466545740764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,10240,0.34453333748711484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,8192,0.0847288899951511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,7168,0.15198311540815565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,8192,0.17837156189812553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,8192,0.33398932880825466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,7168,0.08155822091632418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,6144,0.13254222604963514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,7168,0.06723644336064656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,6144,0.07154133584764269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,7168,0.4300355646345351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,6144,0.06047822369469536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,5120,0.11093778080410427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,5120,0.06254044506284925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,6144,0.33629600207010907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,4096,0.09343022108078003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,5120,0.05574399895138211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,4096,0.058562669489118785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,5120,0.3881813420189752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,3584,0.07924533552593656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,4096,0.04127911064359877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,3584,0.058628446525997586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,4096,0.3303644392225477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,3072,0.0676764448483785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,3584,0.04900888933075798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,3584,0.3311484389834934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,3072,0.05130577749676175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,2560,0.05770044194327461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,3072,0.032404445939593844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,2560,0.044975999328825206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,2560,0.028723554478751287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,2048,0.047354665067460805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,2560,0.32992710007561576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,2048,0.03609600000911289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,2048,0.3231395615471734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,1536,0.03652000096109178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,1536,0.03088444471359253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,3072,0.3607199986775716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,1024,0.025468443830808003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,1536,0.019357333580652874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,1536,0.463880009121365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,1024,0.32978667153252494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,1024,0.015417777829700045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,768,0.019861333900027804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,2048,0.024313777685165405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,768,0.020008888509538438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,512,0.013704000247849358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,768,0.015415110521846347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,768,0.32919732729593915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,512,0.01825066738658481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,512,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,512,0.3146337668100993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,1024,0.022647110952271357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,256,0.015470221638679504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,128,0.006776888751321369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,256,0.010063110954231685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,128,0.014497778481907315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,65536,128,0.010247111320495605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,64,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,128,0.3204133245680067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,32,0.005975111077229182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,64,0.016570667425791424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,65536,32,0.01646844380431705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,65536,256,0.008064889245563084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,65536,0.33626400099860293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,65536,0.17775821685791016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,16384,0.07679288917117648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,65536,256,0.32737776968214244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,16384,0.05585777759552002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,65536,0.1770399941338433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,16384,0.04750844505098131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,16384,0.3136773374345568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,12288,0.04586222105556064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,12288,0.04150399896833632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,65536,0.29130223062303334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,10240,0.03983111182848612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,12288,0.05933244360817803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,10240,0.03183999988767836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,10240,0.3171679973602295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,8192,0.041173332267337374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,8192,0.03534400131967332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,10240,0.050160888168546885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,7168,0.03622488843070136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,8192,0.3223511113060845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,8192,0.028377778000301782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,7168,0.03209155466821458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,6144,0.03253600001335144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,7168,0.025861332813898723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,7168,0.31858576668633354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,6144,0.02913244565327962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,6144,0.02127200033929613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,6144,0.3288951185014513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,5120,0.027196443743175928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,5120,0.02626577847533756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,12288,0.31408267550998265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,4096,0.02311733365058899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,5120,0.33829689025878906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,4096,0.024284443921513025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,4096,0.015830222103330824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,4096,0.33063467343648273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,3584,0.02541422181659275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,3584,0.3277217812008328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,5120,0.021390222840838965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,3072,0.017665778597195942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,3584,0.01757333344883389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,3072,0.022298667165968154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,3072,0.33222489886813694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,2560,0.015490666031837463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,3072,0.013966222604115805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,2560,0.021120000216695998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,2560,0.012352888782819113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,2048,0.013912000589900546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,2048,0.01878666712178124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,3584,0.020351111888885498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,2048,0.3418097760942247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,1536,0.010026666853162978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,1536,0.016934222645229764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,2560,0.34388356738620335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,1536,0.009855999714798397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,1536,0.32560356458028156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,1024,0.008032889001899296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,1024,0.01495911180973053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,1024,0.007988444632954067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,1024,0.3355831040276422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,768,0.007061333292060428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,768,0.013377777404255338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,768,0.3410222265455458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,768,0.007344000041484833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,512,0.005021333280536863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,512,0.012778666284349231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,512,0.32264354493882924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,256,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,512,0.007302222152551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,2048,0.011006222003036074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,256,0.012147555748621622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,128,0.0032337777730491427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,256,0.33691022131178117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,128,0.012111999922328524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,128,0.00666133314371109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,16384,128,0.30728620953030056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,64,0.0031635556370019913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,16384,32,0.0031928889867332247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,64,0.012083555261294046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,16384,32,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,65536,0.24226133028666177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,65536,0.15818933645884195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,16384,256,0.006574222197135289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,16384,0.06465066803826226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,65536,0.15046844217512342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,16384,0.33552177747090656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,12288,0.05192177825503879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,65536,0.3313448958926731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,16384,0.041095998552110456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,12288,0.0414319998688168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,10240,0.045068442821502686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,12288,0.03168888886769613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,16384,0.050456000698937305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,12288,0.3248737653096517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,10240,0.03640000025431315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,8192,0.03780533207787408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,10240,0.027271111806233723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,8192,0.03301599952909682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,7168,0.03274844421280755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,8192,0.023730667101012334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,8192,0.2779573334587945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,7168,0.029765334394243028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,6144,0.028213332096735638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,7168,0.021391999390390184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,7168,0.32552978727552623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,6144,0.027162666122118633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,5120,0.02455111179086897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,6144,0.018604444132910836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,6144,0.3113937907748752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,5120,0.023711111810472276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,4096,0.020939555433061387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,5120,0.3264648914337158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,5120,0.018216000662909616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,4096,0.022481777601771887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,10240,0.40728089544508195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,3584,0.018179555733998615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,4096,0.015053333507643806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,4096,0.31813687748379177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,3584,0.02202577723397149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,3072,0.015799999237060547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,3584,0.33033778932359487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,3072,0.020992888344658747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,2560,0.01407911048995124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,3072,0.012454222473833295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,3072,0.31236089600457084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,2560,0.01940444442960951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,2560,0.3337839974297418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,2560,0.011063111325105032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,2048,0.01721600029203627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,3584,0.013714666995737286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,2048,0.3607333236270481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,1536,0.009715555442704095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,1536,0.01609777741962009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,2048,0.012479999827014076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,1536,0.3307253254784478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,1536,0.00869333330127928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,1024,0.013808888693650564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,2048,0.010085333552625444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,1024,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,1024,0.3320044411553277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,768,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,768,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,768,0.3200017876095242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,768,0.007220444579919179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,512,0.004879111217127906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,1024,0.00790666706032223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,512,0.012791111237472959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,256,0.0039013334446483185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,512,0.006919999917348226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,512,0.3195093207889133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,256,0.012016889121797351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,128,0.0035626664757728577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,256,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,256,0.33285154236687553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,128,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,64,0.00313688897424274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,12288,128,0.006691555596060223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,12288,128,0.29865511258443195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,12288,32,0.003169777699642711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,64,0.01240711079703437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,12288,32,0.011392888923486074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,65536,0.21587110890282524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,65536,0.15174489551120335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,16384,0.05606044663323296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,65536,0.1424515512254503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,16384,0.04817510975731743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,16384,0.03935822182231479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,12288,0.04571555720435249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,16384,0.31093422571818036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,65536,0.3246640099419488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,12288,0.039919111463758684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,10240,0.03958755400445726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,12288,0.3141866789923774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,12288,0.03382044368320041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,10240,0.03576533330811395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,8192,0.030807998445298936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,10240,0.027103111147880554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,10240,0.3117973274654812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,8192,0.031697776582505964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,7168,0.026919111609458923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,8192,0.022263111339675054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,7168,0.028392000330819026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,7168,0.019885333047972787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,7168,0.3288239902920193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,6144,0.024831111232439678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,6144,0.3276568783654107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,8192,0.32622043291727704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,5120,0.0203137778573566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,6144,0.018007111218240526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,5120,0.022663111488024395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,6144,0.023380445109473333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,4096,0.017338666650984023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,5120,0.3310728867848714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,5120,0.016890666551060148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,4096,0.021965333157115515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,3584,0.015469332536061605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,4096,0.01424533294306861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,4096,0.32461510764227974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,3072,0.01313688854376475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,3584,0.0209840006298489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,3584,0.013737777868906656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,3584,0.33322045538160533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,3072,0.019493333167499967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,2560,0.011948444777064852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,3072,0.011944000091817645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,2560,0.01850044396188524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,2048,0.0116462219092581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,2560,0.3097422122955322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,2560,0.010296000374688042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,2048,0.01665599975321028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,1536,0.009022222624884712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,2048,0.009212444225947062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,2048,0.3143093321058485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,1536,0.015048000547620984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,1024,0.007292444507280986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,1536,0.008375999828179678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,1536,0.37400089369879824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,1024,0.013417777915795645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,768,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,1024,0.3040328820546468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,3072,0.33558400472005206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,768,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,512,0.004926222066084544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,768,0.006990222467316522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,768,0.31161867247687447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,512,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,512,0.006664888726340399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,512,0.3423626687791612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,256,0.003863999827040566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,256,0.011855111353927188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,256,0.3488870991600885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,256,0.006339555399285422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,128,0.0032035555276605818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,1024,0.007290666302045186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,128,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,64,0.0031591111587153543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,10240,128,0.30855022536383736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,10240,32,0.0032213332338465583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,64,0.012062222593360476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,10240,32,0.011964444484975604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,65536,0.1591173277960883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,65536,0.10212800237867568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,65536,0.07885066668192546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,10240,128,0.006545777950021956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,16384,0.04829155405362447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,16384,0.03624799847602844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,65536,0.3439759943220351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,12288,0.03203377789921231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,16384,0.02498222225242191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,16384,0.3499751091003418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,12288,0.030808889203601416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,10240,0.02769244379467434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,12288,0.018937778141763475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,12288,0.32602400249905056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,10240,0.027816888358857896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,8192,0.02256977723704444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,10240,0.016961778203646343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,10240,0.3342568874359131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,8192,0.02420711186197069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,7168,0.020626667473051283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,8192,0.015372445185979208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,8192,0.32555288738674587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,7168,0.021885333789719477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,6144,0.018331555856598746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,7168,0.013769778112570444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,7168,0.33273598882887095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,6144,0.02051199972629547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,5120,0.015488000379668342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,6144,0.012793777717484368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,6144,0.32893421914842397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,5120,0.019321777754359774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,4096,0.013556444810496436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,5120,0.012050666742854647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,5120,0.34047823482089573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,4096,0.01904088921017117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,3584,0.0118231111102634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,4096,0.010207110808955299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,4096,0.33956533008151585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,3584,0.018625777628686693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,3072,0.010670222342014313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,3584,0.010035555395815108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,3584,0.33050666915045845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,3072,0.017439999514155917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,2560,0.009728000395827824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,3072,0.00867555538813273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,3072,0.3318471113840739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,2560,0.016249777542220224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,2048,0.009157333109113906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,2560,0.008029333419269985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,2560,0.32771023114522296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,2048,0.015167999598715039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,1536,0.00776800016562144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,2048,0.007310222420427535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,2048,0.32967376708984375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,1536,0.013811555173661975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,1024,0.005607999861240387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,1536,0.0069395556218094295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,1536,0.32387733459472656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,1024,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,768,0.0041617775956789655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,1024,0.006272000157170826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,1024,0.32361777623494464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,768,0.012127999630239276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,512,0.0035537779331207275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,768,0.006324444380071428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,768,0.32293865415785045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,512,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,256,0.0032702222880389956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,512,0.005929777605666055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,512,0.3208906650543213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,256,0.011754666765530905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,128,0.0032239999208185407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,256,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,256,0.3279537889692518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,128,0.011412444214026133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,64,0.0029004443850782183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,8192,128,0.005888000130653381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,8192,128,0.29816532135009766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,8192,32,0.003137777869900068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,64,0.011832888755533429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,65536,0.13963022496965197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,8192,32,0.012079111403889127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,65536,0.09444355302386814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,16384,0.042578667402267456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,65536,0.07613688707351685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,16384,0.03379733363787333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,16384,0.023016888234350417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,12288,0.03199288911289639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,16384,0.4723217752244737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,65536,0.33871645397610134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,12288,0.02954933378431532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,10240,0.027783112393485174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,12288,0.017906667457686532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,12288,0.42129866282145184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,10240,0.026038222842746313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,8192,0.02358755634890662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,10240,0.016141333513789706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,10240,0.3375404410892063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,8192,0.022720000810093347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,7168,0.01865688959757487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,8192,0.014948444234000312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,7168,0.020752888586786058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,8192,0.3729146586524115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,6144,0.01695999999841054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,7168,0.32680267757839626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,6144,0.01937599976857503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,6144,0.33130044407314724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,5120,0.014083556003040738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,5120,0.018594667315483093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,5120,0.3424106703864203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,7168,0.013119999733236102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,4096,0.013087111214796701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,5120,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,4096,0.018421333697107103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,6144,0.012156444291273752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,3584,0.011095110740926532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,4096,0.009503111243247986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,4096,0.34384799003601074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,3584,0.01735733283890618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,3072,0.010051555931568146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,3584,0.009016888837019602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,3584,0.3399377663930257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,3072,0.016521778371598985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,3072,0.008356444537639618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,3072,0.3327217896779378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,2560,0.015817777978049386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,2560,0.3294080098470052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,2560,0.007714666426181793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,2048,0.00835111074977451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,2048,0.014466666513019137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,2048,0.32500534587436253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,2048,0.007361777954631382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,1536,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,1536,0.013532444834709167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,2560,0.00908088849650489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,1024,0.005260444349712796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,1536,0.3191528850131565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,1536,0.006695111178689533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,1024,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,768,0.004249777644872665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,1024,0.006286222073766921
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,1024,0.3180888758765327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,768,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,512,0.003867555409669876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,768,0.006029333505365584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,768,0.3328675429026286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,512,0.012083555261294046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,256,0.00350577798154619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,512,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,512,0.3298479980892605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,256,0.011456888582971362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,128,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,256,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,256,0.33097510867648655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,128,0.010724444356229572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,64,0.0028159999185138275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,7168,128,0.3138871192932129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,7168,32,0.0028880000528362063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,64,0.011036444041464063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,65536,0.11443200376298691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,7168,32,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,65536,0.08663911289638943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,65536,0.07269866598976983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,7168,128,0.005896889087226655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,16384,0.034528887934154935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,65536,0.3358737892574734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,12288,0.028301331732008193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,16384,0.03225244416130914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,16384,0.021872000561820135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,16384,0.3310115602281359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,12288,0.027804444233576458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,10240,0.024775111012988623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,12288,0.3274702231089274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,10240,0.023583110835817125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,8192,0.020984888076782227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,10240,0.01587466730011834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,10240,0.3184106614854601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,8192,0.02198933396074507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,7168,0.01900622248649597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,8192,0.014090667168299357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,8192,0.32611287964714897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,7168,0.01995822290579478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,6144,0.018210666047202218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,7168,0.01291822228166792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,7168,0.3218346701727973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,6144,0.018952000472280715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,12288,0.01759022143152025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,5120,0.015463110473420886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,6144,0.3403280046251085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,6144,0.01183199965291553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,5120,0.017865777015686035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,4096,0.011896000140243106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,5120,0.011039111349317761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,5120,0.33679644266764325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,4096,0.017652443713612027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,3584,0.01018666641579734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,4096,0.009256000320116678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,4096,0.3407786687215169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,3584,0.01676266723208957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,3072,0.009333333207501305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,3584,0.008628444539176093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,3584,0.3324222299787733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,3072,0.015842666228612263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,2560,0.008718222379684448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,3072,0.008280889027648503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,3072,0.3290453222062853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,2560,0.014924445086055331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,2048,0.008364444805516137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,2560,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,2560,0.33168800671895343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,2048,0.013816888961527081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,1536,0.006694222076071634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,2048,0.007251555720965068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,2048,0.33360711733500165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,1536,0.013432888521088494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,1024,0.004920888692140579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,1536,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,1536,0.32848532994588214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,1024,0.012561777399645912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,768,0.0041964443193541635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,1024,0.006355555521117316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,1024,0.3379342291090224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,768,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,512,0.003914666672547658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,768,0.006048000107208888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,768,0.33516799079047305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,512,0.012063110868136087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,256,0.003516444315512975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,512,0.005971555494599872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,512,0.32776443163553876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,256,0.011707555916574268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,128,0.0032239999208185407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,256,0.00563733321097162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,256,0.3334391117095947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,128,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,64,0.002862222285734283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,6144,128,0.005912000106440649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,6144,128,0.3038097752465142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,64,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,6144,32,0.01181244436237547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,65536,0.1039759980307685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,65536,0.08090488778220283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,65536,0.07010400295257568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,16384,0.03161866797341241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,65536,0.3407768938276503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,6144,32,0.0029022221763928733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,16384,0.030233777231640283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,12288,0.025998221503363714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,16384,0.021033777130974665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,16384,0.3207635614607069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,12288,0.025069332785076563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,10240,0.02278577784697215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,12288,0.016927111479971144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,12288,0.3228000005086263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,10240,0.02198222279548645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,8192,0.019884443945354886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,10240,0.015398222539159985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,10240,0.3313902219136556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,8192,0.021091555555661518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,7168,0.017403556240929496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,8192,0.013764444324705334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,8192,0.33107111189100474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,7168,0.01926577753490872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,6144,0.015840889679061044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,7168,0.012455999851226807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,7168,0.29126667976379395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,6144,0.018509333332379658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,5120,0.01388088862101237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,6144,0.01143022212717268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,6144,0.33008355564541286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,5120,0.017550221747822232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,4096,0.01110311100880305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,5120,0.00998133338159985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,5120,0.3167075581020779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,4096,0.017253332667880587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,3584,0.009731555150614845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,4096,0.008796444369686974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,4096,0.3333866596221924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,3584,0.016155555844306946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,3072,0.008958222137557136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,3584,0.008367111285527548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,3584,0.3248364395565457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,3072,0.01517511076397366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,2560,0.00830488900343577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,3072,0.007692444655630324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,3072,0.32299910651312935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,2560,0.014127110441525778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,2048,0.007912000020345053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,2560,0.007391999993059371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,2560,0.3200284375084771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,2048,0.013565333353148567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,1536,0.005967999911970562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,2048,0.006984000404675801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,1536,0.013088888592190213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,2048,0.4538133409288194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,1536,0.006582222051090664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,1536,0.32677777608235675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,1024,0.012459555433856117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,1024,0.31852976481119794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,1024,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,768,0.012105777859687805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,768,0.3352355427212185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,1024,0.004912000149488449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,512,0.003960888832807541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,768,0.006004444426960415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,512,0.01183199965291553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,768,0.004552888787455029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,512,0.3365022341410319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,512,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,256,0.011740444435013665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,256,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,256,0.379040002822876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,128,0.00282577777074443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,128,0.010799999866220685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,5120,128,0.3123928970760769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,64,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,5120,128,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,32,0.002945777856641345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,5120,256,0.0035048888789282907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,64,0.011749332977665795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,65536,0.08583111233181423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,5120,32,0.01071733319097095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,65536,0.07730044258965386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,16384,0.02625511089960734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,65536,0.06731555859247844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,16384,0.026950221922662523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,65536,0.33229509989420575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,16384,0.34070221583048504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,12288,0.02106133268939124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,16384,0.02019466625319587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,12288,0.016550223032633465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,12288,0.0232933329211341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,10240,0.01792355544037289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,12288,0.3192942142486572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,10240,0.020896888441509668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,8192,0.016528000434239704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,10240,0.01482399966981676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,10240,0.3197893301645915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,8192,0.020289777053727042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,8192,0.013070222404268054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,7168,0.014109333356221518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,8192,0.326873779296875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,7168,0.018551111221313477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,6144,0.013363555901580386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,7168,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,7168,0.34337155024210614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,6144,0.017551110850440133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,5120,0.012243555651770698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,6144,0.010378667049937779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,6144,0.3285093307495117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,5120,0.01683999929163191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,4096,0.01075644459989336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,5120,0.009721777505344814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,5120,0.32839645279778373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,4096,0.01549600064754486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,3584,0.009552888572216034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,4096,0.32200532489352757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,4096,0.007967111137178209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,3584,0.01479555500878228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,3072,0.00869955536392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,3584,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,3584,0.32339289453294545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,3072,0.014172444740931192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,2560,0.00812000036239624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,3072,0.007652444144090016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,3072,0.3266471227010091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,2560,0.01385600037044949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,2048,0.006339555399285422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,2560,0.007293333609898885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,2560,0.32524087693956166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,2048,0.013101333545313941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,1536,0.004930666751331753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,2048,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,2048,0.30767999755011666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,1536,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,1024,0.003909333298603694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,1536,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,1536,0.3154924445682102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,1024,0.012409778104888068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,768,0.003946666502290302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,1024,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,768,0.012050666742854647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,512,0.003581333491537306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,768,0.31712267133924693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,1024,0.47929954528808594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,512,0.012118221984969245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,512,0.005635555419656966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,512,0.3215128845638699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,256,0.003216888962520493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,256,0.01166488892502255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,256,0.33365776803758407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,256,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,128,0.0031582222630580268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,768,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,128,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,64,0.0028488888508743713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,4096,128,0.2989822228749593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,4096,128,0.005579555614127054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,4096,32,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,64,0.011487111449241638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,65536,0.0746284458372328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,4096,32,0.010410666465759277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,65536,0.07464622126685248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,16384,0.023306666149033442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,65536,0.06615377797020806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,16384,0.026156443688604567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,65536,0.33454399638705784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,12288,0.018710222509172227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,16384,0.33749156528049046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,16384,0.01995644470055898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,12288,0.02299733294381036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,10240,0.016536888149049547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,12288,0.3204595512813992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,10240,0.02065333392884996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,10240,0.31801244947645396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,8192,0.014118222726715935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,8192,0.0195777780479855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,12288,0.016504888733228047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,8192,0.3289155430263943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,7168,0.013467555244763693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,8192,0.012512000070677863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,7168,0.01816444430086348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,10240,0.014811555544535318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,6144,0.012458666331238218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,7168,0.011333333121405708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,7168,0.334101332558526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,6144,0.017234666479958426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,5120,0.011990221838156382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,6144,0.32513777414957684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,5120,0.016452444924248587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,5120,0.009010666774378883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,5120,0.32587554719713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,4096,0.010445333189434476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,4096,0.0148053334818946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,4096,0.33176978429158527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,4096,0.00870488915178511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,3584,0.009630222287442949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,6144,0.010652444428867765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,3584,0.0147706667582194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,3072,0.00830844458606508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,3584,0.008062221937709385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,3584,0.3390062120225694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,3072,0.013768889009952545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,2560,0.007272000114123027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,3072,0.007616000043021307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,3072,0.3431226677364773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,2560,0.013430222041077085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,2048,0.005931555396980709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,2560,0.007341333561473423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,2048,0.012824888858530255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,2048,0.0069413334131240845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,2048,0.3305697706010607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,1536,0.004898666507667965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,1536,0.012806221842765808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,1536,0.33328265613979763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,1024,0.003983111017280155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,1024,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,2560,0.32588089836968315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,1024,0.006243555496136348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,768,0.0041244443919923575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,1024,0.3316906558142768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,768,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,512,0.0034977777136696708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,768,0.3188720014360216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,512,0.012025777664449481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,1536,0.006678222368160884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,256,0.0031511110977994073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,512,0.005660444498062134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,512,0.31447021166483563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,256,0.011729777687125735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,128,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,256,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,256,0.3260720041063097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,128,0.005583111196756363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3584,128,0.3016284571753608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,64,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3584,32,0.0028355556229750314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,64,0.01106844428512785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3584,768,0.006068444500366847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,32,0.01146666705608368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,65536,0.06144622299406263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,65536,0.07277688715193006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,65536,0.06539822287029691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,16384,0.025415110919210646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3584,128,0.01146755533085929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,16384,0.31869599554273814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,16384,0.019578667150603402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,12288,0.017255110873116385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,16384,0.02035999960369534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,12288,0.022061333060264587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,65536,0.33977333704630536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,12288,0.015847999188635085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,10240,0.014871110518773397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,12288,0.3114382161034478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,10240,0.01995111174053616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,8192,0.013192888763215808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,10240,0.014329777823554145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,10240,0.33156000243292916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,8192,0.018917333748605516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,8192,0.012089778151777057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,7168,0.012444444828563266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,8192,0.325088898340861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,7168,0.018015111486117046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,6144,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,7168,0.33155377705891925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,6144,0.01720711092154185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,6144,0.00979733301533593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,5120,0.01075644459989336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,5120,0.015783111254374187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,7168,0.010382222632567087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,5120,0.00906666699382994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,4096,0.009418666362762451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,6144,0.2759537696838379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,4096,0.014511111709806653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,4096,0.2935386763678657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,3584,0.008056888977686564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,3584,0.014142221874660917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,5120,0.3329457706875271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,3584,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,3072,0.007635555333561367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,3584,0.3119831085205078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,3072,0.01385155568520228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,3072,0.31178222762213814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,3072,0.007687110867765214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,2560,0.006587555425034628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,4096,0.008302222523424361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,2560,0.013439111411571503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,2048,0.005943111247486538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,2560,0.3153902159796821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,2048,0.013117333253224691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,2048,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,2048,0.3077013227674696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,1536,0.005123555660247803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,1536,0.012431110772821637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,1536,0.34278133180406356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,1024,0.0038924444880750445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,1536,0.006266666783226862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,2560,0.007305777735180325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,1024,0.012395555774370829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,768,0.003565333369705412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,1024,0.3190515571170383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,1024,0.006249777972698212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,768,0.01201244443655014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,512,0.0035253332720862497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,768,0.005956444475385878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,768,0.32077688641018337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,512,0.012096888489193387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,256,0.003175111073586676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,512,0.005606222069925732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,512,0.31611111429002553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,256,0.011740444435013665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,128,0.0028862222615215513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,256,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,256,0.31541421678331166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,128,0.011672000090281168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,3072,32,0.002867555452717675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,3072,128,0.3018755647871229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,65536,0.05481955409049988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,32,0.011772444678677453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,3072,64,0.011595555477672152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,65536,0.07043466965357463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,16384,0.018882667024930317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,65536,0.06392266352971394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,65536,0.33251110712687176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,16384,0.025054223007626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,12288,0.01575199928548601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,16384,0.019564444820086162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,16384,0.40348177485995823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,12288,0.021689777572949726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,10240,0.013743999931547375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,3072,128,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,12288,0.015589333242840238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,12288,0.32267822159661186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,10240,0.01996799972322252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,8192,0.011912888950771756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,10240,0.013015111287434896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,10240,0.2991848786671956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,7168,0.011040888726711273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,8192,0.011036444041464063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,8192,0.3201813432905409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,7168,0.017431111799346078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,6144,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,7168,0.010371555884679159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,7168,0.34811822573343915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,6144,0.016232000456915963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,6144,0.32716443803575307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,6144,0.009675555759006077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,5120,0.009678222239017487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,5120,0.015154666370815702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,5120,0.3176533381144206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,8192,0.018944889307022095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,5120,0.009000889129108852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,4096,0.014160888062583076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,4096,0.32515109909905326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,4096,0.00831288927131229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,3584,0.008063111040327284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,3584,0.013821333646774292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,3584,0.32548711034986705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,3072,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,4096,0.008665777742862701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,3072,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,3072,0.007319110963079665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,2560,0.006253333141406377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,3072,0.300494220521715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,2560,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,2560,0.3004666699303521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,2560,0.006958222223652734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,2048,0.005231110999981563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,3584,0.008064000142945183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,2048,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,1536,0.004598222259018156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,2048,0.0069653333889113525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,2048,0.3110906547970242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,1536,0.012431999875439538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,1536,0.00665155549844106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,1536,0.34164799584282773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,1024,0.01242844429281023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,1024,0.006275555739800136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,1024,0.32673599984910756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,768,0.0038515557017591265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,768,0.011981333295504252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,768,0.33300712373521596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,768,0.005942222144868638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,512,0.003528000166018804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,1024,0.003924444317817688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,512,0.011796444654464722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,256,0.0031644445326593188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,512,0.3262551095750597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,512,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,256,0.01144088887506061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,128,0.0029048888633648553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,256,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,256,0.3229288789961073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,128,0.011376000112957425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,64,0.0028080000645584534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2560,128,0.005647999958859549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2560,128,0.30578311284383136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,64,0.011406222151385413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,65536,0.0449022220240699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2560,32,0.010403555300500657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,65536,0.06699822346369426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,65536,0.32604000303480357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,65536,0.0621324446466234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,16384,0.01648977730009291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,16384,0.02464444438616435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2560,32,0.002862222285734283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,16384,0.018927110566033255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,12288,0.021104888783560857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,12288,0.331657780541314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,12288,0.014327110515700446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,10240,0.012227555943859948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,16384,0.32577866978115505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,10240,0.019896888070636325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,12288,0.013358222113715278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,8192,0.012816888590653738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,10240,0.013085333009560903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,8192,0.018206222189797294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,7168,0.012310222619109683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,10240,0.46362935172186953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,8192,0.0107004443804423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,8192,0.31829955842759877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,6144,0.01127466642194324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,7168,0.016600888636377122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,7168,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,6144,0.015878222054905362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,5120,0.010335110955768162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,6144,0.0096142225795322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,7168,0.41040176815456814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,6144,0.3350959883795844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,4096,0.009280000295903947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,5120,0.014841778410805596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,5120,0.009082666701740688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,5120,0.32914490169949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,4096,0.014451555079884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,3584,0.00849955528974533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,4096,0.008033778104517195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,4096,0.3227608998616536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,3584,0.01388711151149538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,3072,0.006694222076071634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,3584,0.007920000288221572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,3584,0.3173466788397895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,3072,0.013451555536852943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,3072,0.007663999994595845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,3072,0.324614233440823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,2560,0.013430222041077085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,2560,0.3077884515126546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,2560,0.007280000381999546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,2048,0.005232000102599462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,2048,0.012938666674825879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,2048,0.31974932882520885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,2048,0.0069439998931354946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,1536,0.012738666600651212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,2560,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,1536,0.0063493334584765965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,1024,0.0041964443193541635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,1536,0.3316391044192844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,1024,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,1024,0.006025777922736273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,768,0.003559999995761447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,1536,0.004528000122971005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,768,0.011765333513418833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,768,0.31185245513916016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,512,0.003579555700222651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,512,0.011770666473441653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,1024,0.3313679960038927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,512,0.3354159990946452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,256,0.0032053333189752367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,512,0.005839111076460944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,256,0.011726222104496427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,768,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,256,0.3173617786831326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,128,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,256,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,2048,128,0.005633777628342311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,64,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,2048,128,0.2796764373779297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,2048,32,0.002833777831660377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,64,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,65536,0.034224887688954674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,32,0.01032977799574534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,65536,0.06608711348639594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,65536,0.3192080126868354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,65536,0.057998220125834145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,2048,128,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,16384,0.31769243876139325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,16384,0.018582221534517076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,12288,0.013581333061059317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,12288,0.02102311121092902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,16384,0.015875554747051664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,12288,0.013752000199423896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,10240,0.012252444194422828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,12288,0.3167537848154704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,10240,0.018812444474962022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,10240,0.29480711619059247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,8192,0.011024889018800525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,10240,0.012177777787049612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,8192,0.017582222819328308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,8192,0.3309600088331434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,7168,0.010654222634103564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,16384,0.02403733299838172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,7168,0.01643733349111345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,7168,0.010031999813185798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,7168,0.33827821413675946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,6144,0.010149333212110732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,6144,0.0158906661801868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,6144,0.33462044927808976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,6144,0.009720889230569204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,8192,0.011335999601417117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,5120,0.01479822231663598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,4096,0.008366222182909647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,5120,0.009103111094898647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,5120,0.30964355998569065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,4096,0.014170666535695394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,3584,0.007381333245171442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,4096,0.3290879991319445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,3584,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,5120,0.009072889056470659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,3584,0.32149243354797363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,3072,0.00638755535085996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,3584,0.008055999875068665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,3072,0.013550221920013428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,2560,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,3072,0.007679111427730984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,3072,0.32056087917751735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,2560,0.01311466677321328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,2048,0.005210666606823604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,2560,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,4096,0.00811644477976693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,2560,0.3732284439934625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,1536,0.004640000147951974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,2048,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,2048,0.006970666348934174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,2048,0.32569066683451336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,1536,0.006589333216349284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,1024,0.0041928887367248535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,1536,0.012470222181744046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,1536,0.3303431140051948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,1024,0.012421333127551608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,768,0.003572444534964032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,1024,0.006223111102978389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,1024,0.3148924509684245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,512,0.003544888976547453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,768,0.005977777971161737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,768,0.3236622280544705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,512,0.011735999749766456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,512,0.35351022084554035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,512,0.005577777822812398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,256,0.0029582221888833572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,256,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,256,0.3420248826344808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,256,0.005550222264395819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,128,0.0029013332807355454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,128,0.0110871113008923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,768,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,64,0.0025271111064487034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1536,128,0.28178487883673775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1536,128,0.005576888720194499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1536,32,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,65536,0.02516888909869724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,64,0.01149866647190518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1536,32,0.01034044474363327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,65536,0.06355555852254231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,16384,0.012502222425407834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,65536,0.05754488706588745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,65536,0.3342657883961995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,16384,0.023249778482649062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,16384,0.016511110795868766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,16384,0.3410337766011556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,12288,0.01958400011062622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,12288,0.013800889253616333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,10240,0.009777777724795872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,10240,0.018203554881943595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,10240,0.32107732031080455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,12288,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,8192,0.00928177767329746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,10240,0.012400888734393649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,8192,0.01650399963061015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,12288,0.3329253461625841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,8192,0.011046222514576383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,8192,0.3293946584065755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,7168,0.0087013335691558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,7168,0.01590844492117564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,7168,0.010057777994208867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,6144,0.007954667011896769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,7168,0.32283467716640896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,6144,0.015200000670221118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,5120,0.0069822221994400024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,6144,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,6144,0.34518933296203613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,5120,0.014449778530332776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,4096,0.007306666837798224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,5120,0.00906577789121204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,5120,0.32640356487698025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,4096,0.013807111316257052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,3584,0.007000889215204451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,4096,0.008309333688682979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,4096,0.3379964563581679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,3584,0.013766222529941134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,3072,0.0064026667839950984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,3584,0.007999111380841997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,3584,0.32364622751871747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,3072,0.01313866674900055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,2560,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,3072,0.007640000018808577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,3072,0.3380151059892442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,2560,0.01275199982855055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,2048,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,2560,0.007289778027269576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,2560,0.3178222179412842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,2048,0.012456888953844706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,1536,0.005223999834722943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,2048,0.006984000404675801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,2048,0.34836355845133465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,1536,0.012435555458068848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,1024,0.0042151109211974675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,1536,0.3160791132185194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,1536,0.006678222368160884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,1024,0.011755555868148804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,768,0.004056888735956616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,1024,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,1024,0.3147119945949978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,768,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,512,0.003576000117593341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,768,0.2836835649278429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,768,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,512,0.011727111207114326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,256,0.003176888864901331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,512,0.30907199117872447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,256,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,256,0.005559999909665849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,256,0.30295467376708984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,128,0.01146311147345437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,128,0.005564444594913059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,64,0.0028728888266616394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,1024,128,0.44743023978339297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,1024,512,0.005922666854328579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,32,0.0032026666320032547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,64,0.011372444530328115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,65536,0.026529777381155226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,1024,32,0.011367999845080905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,65536,0.061295111974080406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,16384,0.011391110718250275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,65536,0.057612445619371205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,1024,128,0.0031946665710873077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,65536,0.4172506597306993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,16384,0.02195022172398037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,12288,0.009420444567998251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,16384,0.3182355562845866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,16384,0.01624000072479248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,12288,0.018931556079122756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,10240,0.008994667066468133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,12288,0.01349244432316886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,12288,0.31829600863986546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,10240,0.017880888448821176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,8192,0.008933333059151968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,10240,0.012430222498046027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,10240,0.38162578476799863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,8192,0.016493333710564505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,7168,0.008287111090289222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,8192,0.010747555229398938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,8192,0.32350044780307347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,7168,0.01590044465329912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,6144,0.007306666837798224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,7168,0.01036088913679123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,7168,0.3602444330851237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,6144,0.33120621575249565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,6144,0.009668444593747457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,5120,0.006968888971540663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,5120,0.014883556299739413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,5120,0.324760013156467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,4096,0.007695111135641734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,5120,0.00906755526860555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,4096,0.014096889230940076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,4096,0.3583964506785075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,6144,0.01516266663869222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,3584,0.0070426662762959795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,3584,0.013802666631009845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,3584,0.34988445705837673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,3584,0.00794044468137953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,3072,0.007101332975758447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,3072,0.013189333180586496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,3072,0.34299198786417645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,3072,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,2560,0.006673777682913675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,2560,0.01309511148267322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,2560,0.32862043380737305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,2560,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,2048,0.005312888986534542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,2048,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,2048,0.3286222351921929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,2048,0.006976000136799282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,1536,0.004493333399295807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,1536,0.012510221865442065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,1536,0.3253822326660156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,1024,0.003911111089918348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,4096,0.008296888735559251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,1024,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,1024,0.30380090077718097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,1024,0.006330666856633292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,768,0.011807111402352652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,768,0.3403280046251085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,1536,0.006650666809744305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,512,0.0032595555401510666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,768,0.005942222144868638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,512,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,768,0.003519111209445529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,512,0.005655111124118169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,256,0.003157333367400699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,256,0.011456888582971362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,256,0.30168178346421987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,128,0.003198222153716617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,128,0.011791110866599612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,128,0.2816319995456272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,128,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,64,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,64,0.011589333415031433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,768,32,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,768,256,0.005616000129116907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,65536,0.01966399947802226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,768,32,0.010407110883129967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,65536,0.06024977895948622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,16384,0.009711999860074785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,65536,0.3093493249681261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,65536,0.05139644609557258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,16384,0.021418665846188862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,12288,0.008635555704434713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,16384,0.30407823456658256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,768,512,0.3387448787689209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,16384,0.016215110818545025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,12288,0.018894222047593858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,12288,0.013478221992651621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,12288,0.30240710576375324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,10240,0.01772799922360314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,8192,0.007975999679830339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,10240,0.0123848890264829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,10240,0.33055822054545086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,8192,0.016544888416926067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,7168,0.00794933322403166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,8192,0.010827555424637265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,8192,0.30825334125094944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,7168,0.01587022178702884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,6144,0.007318221860461765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,7168,0.010395555860466426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,7168,0.32699110772874623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,6144,0.01514488955338796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,10240,0.007737778127193451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,5120,0.006984000404675801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,6144,0.009691555466916826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,6144,0.320067564646403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,5120,0.014490667316648694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,5120,0.009008000294367472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,4096,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,5120,0.3318657875061035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,4096,0.014144000079896716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,3584,0.0069919998447100324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,4096,0.008096000386608971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,4096,0.3630355464087592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,3584,0.013825777504179211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,3072,0.0070088886552386815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,3584,0.00776800016562144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,3584,0.3375679916805691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,3072,0.013436444103717804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,2560,0.006931555353932911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,3072,0.007317333585686154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,3072,0.3579440116882324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,2560,0.013443555268976422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,2048,0.006236444330877728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,2560,0.006994666324721442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,2560,0.33021510971917045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,2048,0.012805333567990197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,1536,0.00453599997692638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,2048,0.0069919998447100324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,2048,0.3425288995107015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,1536,0.33096088303460014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,1536,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,1024,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,1024,0.011913778053389655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,1024,0.31416355239020455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,1024,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,768,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,1536,0.01276088919904497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,768,0.34341245227389866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,768,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,512,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,512,0.011968000067604912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,512,0.005910222315125995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,768,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,256,0.0031555555760860443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,256,0.011427555647161273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,256,0.3304159906175402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,256,0.005576888720194499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,128,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,128,0.3084639973110623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,512,512,0.3366364373101129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,64,0.0027991111079851785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,512,128,0.005444444302055571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,32,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,64,0.010729778144094678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,65536,0.014782221780882942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,512,32,0.011021333436171213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,512,128,0.0028719999310043124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,65536,0.05923200315899319
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,16384,0.008831111093362173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,65536,0.049979554282294385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,65536,0.33115821414523655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,16384,0.02093688812520769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,12288,0.00833422193924586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,16384,0.0165964447789722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,16384,0.3120533360375298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,12288,0.018582221534517076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,12288,0.013783111340469785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,12288,0.32625865936279297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,10240,0.018035555879275005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,10240,0.0124195557501581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,10240,0.3204409016503228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,8192,0.007297777467303806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,8192,0.016540444559521146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,8192,0.32478756374782985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,8192,0.01074577785200543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,7168,0.006892444358931647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,10240,0.007626666790909237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,7168,0.3246462345123291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,6144,0.006697777658700943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,7168,0.010063110954231685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,6144,0.01550222271018558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,5120,0.006258666515350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,6144,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,5120,0.01483022173245748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,6144,0.47066757414076066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,4096,0.006027555714050929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,5120,0.29738577206929523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,5120,0.009000889129108852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,4096,0.013857777747843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,7168,0.01587199999226464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,3584,0.006728000111050076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,4096,0.300413343641493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,3584,0.007640889121426477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,3072,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,3584,0.391175111134847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,3072,0.013105777402718862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,3072,0.007263110743628607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,2560,0.006353777729802662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,3072,0.3653822210099962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,2560,0.012748444245921241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,4096,0.007982222570313348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,2560,0.3577253288692898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,2560,0.007283555964628856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,2048,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,3584,0.01369599997997284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,2048,0.01313688854376475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,2048,0.00692622239391009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,1536,0.004325333154863781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,2048,0.3454097906748454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,1536,0.012484444512261285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,1024,0.003918222255176968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,1536,0.006660444455014334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,1536,0.31473689609103733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,1024,0.012242666549152799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,1024,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,768,0.0035822221802340615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,1024,0.3416871229807536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,768,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,512,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,768,0.0059573331640826324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,512,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,512,0.005802666561471091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,512,0.34418667687310117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,256,0.0031715554909573663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,256,0.01164177805185318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,256,0.3304426670074463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,128,0.0029164445069101122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,256,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,128,0.011747555600272285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,768,0.34075289302402073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,64,0.0028497777465316984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,256,128,0.31942134433322483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,256,128,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,64,0.011471110913488599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,65536,0.014211555322011312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,256,32,0.011378666592968835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,65536,0.05581511391533745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,65536,0.33519376648796934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,16384,0.008025777836640676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,65536,0.048651556173960366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,16384,0.021724444296624925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,16384,0.34456266297234434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,256,32,0.0027546667390399506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,16384,0.016566221912701923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,12288,0.019000000423855253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,12288,0.34006577067905003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,12288,0.013449778159459433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,10240,0.007296889192528195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,10240,0.01791288952032725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,10240,0.3267209000057644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,8192,0.006948444578382704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,10240,0.01239466667175293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,8192,0.016794666647911072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,12288,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,7168,0.006676444576846228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,8192,0.3306159973144531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,8192,0.011064888702498542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,7168,0.01586577792962392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,6144,0.006592888798978593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,7168,0.010334221853150262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,7168,0.31211556328667533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,6144,0.01514488955338796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,5120,0.0063440000845326324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,6144,0.009671111073758867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,6144,0.32513777414957684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,5120,0.014518222875065274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,4096,0.006025777922736273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,5120,0.009009777671760982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,5120,0.31843021180894643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,4096,0.014069333672523499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,3584,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,4096,0.008068444828192392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,4096,0.3238897853427463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,3584,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,3584,0.007887110941939885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,3072,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,3584,0.3235386742485894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,3072,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,3072,0.007598222129874759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,3072,0.33895111083984375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,2560,0.013076444466908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,2560,0.007245333658324347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,2560,0.32751467492845326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,2048,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,2048,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,2048,0.3313475449879964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,2048,0.0069617778062820435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,1536,0.004556444370084339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,1536,0.012461333639091916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,2560,0.006263111200597551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,1024,0.0038568890757030914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,1536,0.32874488830566406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,1536,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,768,0.003705777641799715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,1024,0.012463111016485425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,1024,0.006128888991143968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,768,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,512,0.0034942221310403612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,768,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,768,0.3117706775665283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,1024,0.4677537812127008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,512,0.011702222128709158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,256,0.0029475556479560006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,512,0.32577599419487846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,512,0.00565155554148886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,256,0.011481777661376528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,128,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,256,0.0053431110249625305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,256,0.3237413300408257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,128,0.011367110742463006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,64,0.0028773333049482773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,32,128,128,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,128,32,0.0029022221763928733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,32,128,128,0.3595297866397434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,64,0.01108266661564509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,65536,0.010749333434634738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,16384,0.006999111009968652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,128,32,0.011069333387745751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,12288,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,16384,0.021008888880411785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,10240,0.005978666659858491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,12288,0.01889333294497596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,8192,0.00657155571712388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,10240,0.01722133325205909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,7168,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,8192,0.016527111331621807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,6144,0.006589333216349284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,7168,0.01548266741964552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,5120,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,6144,0.015454222758611044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,4096,0.005947555518812603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,5120,0.014164444473054675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,65536,0.05517066518465678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,4096,0.014165333575672574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,3072,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,3584,0.0137155552705129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,2560,0.006265777680608962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,3072,0.01350400017367469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,2048,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,2560,0.01312355531586541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,1536,0.004323555363549126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,2048,0.012819555898507437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,1024,0.0039022221333450745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,1536,0.012471111284361945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,768,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,1024,0.012111110819710625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,512,0.003565333369705412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,768,0.011781333221329583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,3584,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,256,0.0031644445326593188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,512,0.011740444435013665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,128,0.0028239999794297745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,64,0.002899555489420891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,256,0.011035555766688453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,128,0.011408888631396823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,64,32,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,64,0.011366222467687396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,65536,0.008991111483838823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,16384,0.006575999988449945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,64,32,0.0110871113008923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,65536,0.05494399865468343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,16384,0.021006221572558086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,10240,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,12288,0.018696889281272888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,8192,0.005958222266700532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,10240,0.01721155477894677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,7168,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,8192,0.016548444827397663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,6144,0.006272888845867581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,7168,0.015490666031837463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,5120,0.005943999936183293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,6144,0.0151235560576121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,4096,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,5120,0.014140443669425117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,3584,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,4096,0.014103111293580798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,12288,0.006342222293217977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,3072,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,2560,0.005974222388532426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,3584,0.013420444395807056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,2048,0.005290666802061929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,3072,0.013407111167907715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,1536,0.004541333350870344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,2560,0.013131555583741931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,1024,0.0038444445365005066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,2048,0.012426666915416718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,768,0.0032497776879204642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,1536,0.012439111040698158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,512,0.0032302221904198327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,1024,0.012155555188655853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,256,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,768,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,128,0.0025759999536805684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,256,0.011378666592968835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,64,0.002543111021320025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,128,0.010378667049937779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,32,32,32,0.002542222125662698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,64,0.01145155562294854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,32,0.010027555955780877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,16384,0.30171023474799263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,16384,0.1626933283276028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,32,32,512,0.011718221836619906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,16384,0.1447804503970676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,12288,0.2181208928426107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,12288,0.12858133845859104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,16384,0.3419022295210097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,10240,0.18381244606441924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,12288,0.1099662250942654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,10240,0.10754844215181138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,12288,0.34586132897271055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,10240,0.10218933555814956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,8192,0.14710844887627497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,8192,0.09383377763960098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,10240,0.33788087632921004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,7168,0.12828089131249323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,8192,0.07637511359320746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,7168,0.10351022084554036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,8192,0.32845955424838597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,6144,0.11174489392174615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,7168,0.0673475530412462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,6144,0.09799466530481975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,7168,0.38862933052910703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,5120,0.09348533550898235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,6144,0.05877421961890327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,6144,0.3379608790079753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,5120,0.0838586688041687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,4096,0.07865777942869398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,5120,0.04961244596375359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,5120,0.34616178936428493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,4096,0.040942221879959106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,3584,0.06700533628463745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,4096,0.2927795516120063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,3584,0.0647066699133979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,3584,0.036728888750076294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,3584,0.30109956529405385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,4096,0.06515466504626803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,3072,0.05122399992412991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,3072,0.35748889711168075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,2560,0.049230221245023936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,2560,0.053103999959097974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,3072,0.05814399984147814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,2560,0.028115553988350764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,2560,0.2881431049770779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,2048,0.04454666707250807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,2048,0.04053599966896905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,2048,0.02400355537732442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,1536,0.03141155507829454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,1536,0.03345155715942383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,2048,0.4622693591647678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,1024,0.022579555710156757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,1536,0.3266879982418484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,1024,0.022662222385406494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,1024,0.3072151078118218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,768,0.01789688898457421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,1024,0.015088000231319003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,3072,0.03228711088498434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,768,0.015236443943447538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,768,0.3839706579844157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,512,0.013073777986897362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,1536,0.019352000620630052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,512,0.018610666195551556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,256,0.00832622249921163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,512,0.31698044141133624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,256,0.015531554818153381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,256,0.30629777908325195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,128,0.006616000086069107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,256,0.010055111514197456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,768,0.022304889228608873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,128,0.013832000394662222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,128,0.010032888915803697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,65536,128,0.33405155605740017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,32,0.005694222119119432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,64,0.015455999308162265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,65536,32,0.016514667206340365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,65536,512,0.012708444562223224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,65536,0.32268532117207843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,65536,0.1687626706229316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,16384,0.08910133441289265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,65536,64,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,16384,0.05432266659206814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,65536,0.18428266048431396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,12288,0.06889155838224623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,16384,0.046471112304263644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,12288,0.044919110006756253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,16384,0.3168506622314453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,12288,0.039319111241234675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,10240,0.059402664502461754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,12288,0.34909510612487793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,65536,0.322197331322564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,10240,0.03944622145758735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,8192,0.041895111401875816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,8192,0.03714311122894287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,10240,0.03344800074895223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,10240,0.30420178837246364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,7168,0.03689777851104736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,8192,0.026113778352737427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,7168,0.04029333260324266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,8192,0.31281065940856934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,6144,0.03189155459403992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,7168,0.02831288841035631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,7168,0.33696622318691677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,6144,0.03774133324623108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,5120,0.02757955590883891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,6144,0.020425778296258714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,6144,0.3207520114050971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,5120,0.034789333740870156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,4096,0.023373333944214716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,5120,0.019711110326978896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,5120,0.3398204379611545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,4096,0.028961777687072754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,3584,0.020638222495714825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,4096,0.3139004442426893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,4096,0.015788444214397006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,3584,0.02632177703910404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,3072,0.01830666595035129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,3584,0.014389332797792224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,3584,0.3147146701812744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,3072,0.023609777291615803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,2560,0.015531554818153381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,3072,0.013920888304710388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,3072,0.33649600876702207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,2560,0.02091377807988061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,2048,0.01587733295228746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,2560,0.012146666646003723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,2560,0.31825778219434947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,2048,0.01899111105336083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,1536,0.010404444403118556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,2048,0.011121778024567498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,2048,0.33586843808492023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,1536,0.01735555628935496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,1024,0.008015111088752747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,1536,0.009389333426952362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,1536,0.32482756508721244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,1024,0.014194667339324951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,768,0.0069475554757648045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,1024,0.009003555609120263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,1024,0.32959643999735516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,768,0.01424533294306861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,512,0.004771555463473002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,768,0.007378666765160031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,768,0.31919556193881565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,512,0.013165333204799227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,256,0.003463111196955045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,512,0.00703999979628457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,512,0.3283582263522678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,256,0.012390221986505719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,128,0.003168888803985384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,256,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,256,0.32321421305338544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,128,0.012124444047609964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,64,0.0029333333174387612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,16384,128,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,16384,32,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,16384,128,0.31380976570977104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,64,0.012117333710193634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,16384,32,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,65536,0.23614133728875053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,65536,0.149144000477261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,16384,0.06285422378116183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,16384,0.048751112487581044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,65536,0.15240800380706787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,16384,0.04101066788037618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,12288,0.05063644382688734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,16384,0.3413360118865967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,65536,0.3025360107421875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,12288,0.04073333409097459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,10240,0.04299822118547228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,12288,0.032617777585983276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,10240,0.035735110441843666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,10240,0.028155555327733357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,8192,0.03617422117127313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,10240,0.3070346779293484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,8192,0.03251466817326016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,8192,0.3045297728644477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,8192,0.022639110684394836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,7168,0.03241689006487528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,7168,0.035903112755881414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,12288,0.33045866754319936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,6144,0.02889333168665568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,7168,0.3111342324150933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,7168,0.02128977742460039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,6144,0.03288977675967746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,6144,0.018357333209779527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,5120,0.024511110451486375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,6144,0.3021155463324653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,5120,0.028855111863878038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,4096,0.021893334057595994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,5120,0.016892444756295946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,5120,0.32996977700127494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,4096,0.02664622167746226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,3584,0.019299555155966017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,4096,0.014183999763594734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,4096,0.31640532281663686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,3584,0.023823110593689814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,3072,0.017162667380438913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,3584,0.013584000368913015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,3584,0.31231289439731175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,2560,0.012287110918098025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,3072,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,3072,0.022058667408095464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,2560,0.019467555814319186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,3072,0.4721493191189236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,2560,0.32617065641615123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,2048,0.012064889073371887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,2048,0.018193778064515855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,2048,0.3133804533216688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,2048,0.009675555759006077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,1536,0.00902133352226681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,1536,0.015819556183285184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,1536,0.3207351101769341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,1024,0.007340444458855524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,1536,0.008714666797055138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,2560,0.011047999891969891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,1024,0.014202665951516895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,768,0.006172444257471297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,1024,0.32521600193447536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,1024,0.00831733312871721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,768,0.0135057775510682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,512,0.0052133335007561584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,768,0.0075839997993575195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,768,0.3220631016625298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,512,0.012798222402731577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,256,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,512,0.0069297779765393995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,512,0.3177635669708252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,256,0.012434666355450949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,128,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,256,0.006569777925809224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,256,0.3254968855116102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,128,0.012068444656001197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,64,0.003162666741344664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,12288,128,0.006360000206364526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,12288,128,0.31146489249335396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,12288,32,0.0032044444233179092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,64,0.012182222472296821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,12288,32,0.012810666528013019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,65536,0.2005386617448595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,65536,0.14411021603478327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,16384,0.05901333358552721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,65536,0.16293599870469835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,16384,0.047712001535627574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,16384,0.03970400161213345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,12288,0.045680890480677284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,16384,0.32668977313571507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,65536,0.33553510242038304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,12288,0.039057777987586126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,10240,0.0381644434399075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,12288,0.029634667767418757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,12288,0.31016355090671116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,10240,0.03495288888613383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,8192,0.0314648879898919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,10240,0.028120001157124836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,10240,0.3363946543799506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,8192,0.030924446052975122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,7168,0.028576887316173975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,8192,0.021475555168257818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,8192,0.30858755111694336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,7168,0.03123733401298523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,6144,0.025913778278562758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,7168,0.020266667008399963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,7168,0.3283199999067518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,6144,0.027646222048335608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,6144,0.01762488815519545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,5120,0.022330666581789654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,6144,0.32000801298353404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,5120,0.0254711111386617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,5120,0.016368889146380954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,4096,0.019352000620630052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,5120,0.334133333630032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,4096,0.02345422241422865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,4096,0.013831111292044321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,3584,0.015579554769727917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,4096,0.33543377452426487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,3584,0.02165777815712823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,3072,0.01389244364367591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,3584,0.012768000364303589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,3584,0.32622400919596356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,3072,0.02001688877741496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,3072,0.011759111450778114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,3072,0.32611555523342556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,2560,0.018684445155991446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,2560,0.3015822304619683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,2560,0.01033688916100396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,2048,0.011382222175598145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,2048,0.016957332690556843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,2048,0.3176204363505046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,2048,0.00904444439543618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,1536,0.00870311094654931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,1536,0.01477866702609592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,2560,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,1024,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,1536,0.31693956587049693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,1536,0.008210666477680206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,1024,0.013471110827393003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,768,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,1024,0.3134480052524143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,768,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,512,0.004557333472702238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,768,0.007311111523045435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,768,0.3010524378882514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,512,0.012410666379663678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,512,0.006983111302057902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,512,0.31570222642686635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,256,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,1024,0.00777511133088006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,256,0.3335004382663303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,128,0.003536888708670934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,128,0.011746666497654386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,256,0.0038648889296584656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,10240,128,0.3080977863735623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,64,0.0031662223239739737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,128,0.006620444357395172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,10240,32,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,64,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,10240,32,0.012854222622182635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,65536,0.15664800008138022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,10240,256,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,65536,0.09396000040902032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,16384,0.04658577839533488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,65536,0.0803955528471205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,16384,0.03487733337614272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,16384,0.02478133307562934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,12288,0.03199822372860379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,16384,0.3347288767496745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,65536,0.3435315555996365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,12288,0.029743999242782593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,10240,0.02734666731622484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,12288,0.018719999326599967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,12288,0.32461510764227974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,10240,0.02701777716477712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,10240,0.018087999688254464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,8192,0.022138666775491502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,10240,0.3341395590040419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,8192,0.02292622294690874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,7168,0.02019999921321869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,8192,0.015159999330838522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,7168,0.025739555557568867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,7168,0.013802666631009845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,7168,0.34124088287353516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,6144,0.018025777406162687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,6144,0.023376888699001733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,6144,0.33100443416171604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,6144,0.012453333371215396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,5120,0.015445333388116626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,5120,0.022077333596017625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,8192,0.3323582278357612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,4096,0.013739555246300168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,5120,0.32682932747734916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,5120,0.011423110961914062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,4096,0.01993777851263682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,3584,0.012004444168673621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,4096,0.32042309972974986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,3584,0.01887999971707662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,3584,0.009358222285906473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,3584,0.31998756196763783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,3072,0.01051644484202067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,3072,0.01736177835199568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,3072,0.33844354417588973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,3072,0.0086666668454806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,2560,0.009659555223253038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,4096,0.010042666561073726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,2560,0.01627466744846768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,2048,0.00907466643386417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,2560,0.31262845463222927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,2560,0.007703999678293864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,2048,0.014922666880819531
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,1536,0.007616000043021307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,2048,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,2048,0.3088880115085178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,1536,0.013457777599493662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,1024,0.005633777628342311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,1536,0.007304000357786815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,1536,0.3223857879638672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,1024,0.012473777764373355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,768,0.00423555572827657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,1024,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,1024,0.2978915638393826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,768,0.012431110772821637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,512,0.0038782221575578055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,768,0.006574222197135289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,768,0.3232053385840522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,256,0.0035422220826148987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,512,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,512,0.01235466698805491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,512,0.32637066311306423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,256,0.011751111182901593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,128,0.003030222323205736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,256,0.2948933442433675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,256,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,128,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,64,0.0029395555870400537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,8192,32,0.0028924445311228433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,8192,128,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,8192,128,0.2972337669796414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,64,0.011387555963463254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,65536,0.13881422413720026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,8192,32,0.012117333710193634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,16384,0.039015110996034406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,65536,0.08592710892359416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,16384,0.03288711110750834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,16384,0.3332044548458523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,65536,0.3466773298051622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,16384,0.02251200046804216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,12288,0.03093155556254917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,12288,0.028194665908813477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,12288,0.017625777257813346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,12288,0.3380257818433974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,10240,0.027063111464182537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,65536,0.07618400123384264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,10240,0.024364444944593642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,8192,0.023200889428456623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,10240,0.016785777277416654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,10240,0.3329777717590332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,8192,0.022337777747048274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,7168,0.020301333732075162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,8192,0.014564444621404013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,8192,0.3188737763298882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,7168,0.024136889312002394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,6144,0.016883555385801528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,7168,0.01313333296113544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,7168,0.33893778589036727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,6144,0.022724444667498272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,5120,0.014377777775128683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,6144,0.012062222593360476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,6144,0.319851557413737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,5120,0.02159822152720557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,4096,0.012772444221708508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,5120,0.011260444091426002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,5120,0.3281102180480957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,4096,0.01957155598534478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,3584,0.011156444748242697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,4096,0.009630222287442949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,4096,0.3219208982255724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,3584,0.018205333087179396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,3072,0.010144000252087912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,3584,0.3345902231004503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,3072,0.01718577742576599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,3072,0.008207999997668797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,3072,0.3259306748708089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,2560,0.016160888804329764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,2560,0.3127342330084907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,3584,0.008739555875460306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,2048,0.008664888640244802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,2560,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,2048,0.014800000521871777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,2048,0.007322666545708974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,2048,0.32035644849141437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,1536,0.007317333585686154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,1536,0.013533333109484779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,1536,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,1024,0.005344888816277186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,1024,0.012488000094890594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,2560,0.009087111386987898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,1024,0.3222506576114231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,768,0.004221333397759331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,1024,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,768,0.012489777472284106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,1536,0.34532621171739364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,768,0.005931555396980709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,512,0.003593777616818746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,768,0.3434586789872911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,512,0.012058667010731168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,256,0.003286222202910317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,512,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,512,0.3309271070692274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,256,0.011049778097205691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,128,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,256,0.005920889063013925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,256,0.3320675426059299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,128,0.0107351111041175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,7168,128,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,7168,128,0.2966710991329617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,32,0.0029084444459941653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,64,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,7168,32,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,65536,0.12545600202348497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,65536,0.07928533024258085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,16384,0.034536888202031456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,65536,0.0744346645143297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,7168,64,0.0028853333658642242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,16384,0.03119822343190511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,65536,0.34382043944464785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,12288,0.02787644333309597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,16384,0.02162666618824005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,16384,0.3277084562513563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,12288,0.026717333330048457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,12288,0.017287110288937885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,12288,0.32593954934014213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,10240,0.02274488906065623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,8192,0.020954666866196528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,10240,0.016196444630622864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,10240,0.3266879982418484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,8192,0.021643555826610986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,7168,0.01870844430393643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,8192,0.013879111243618859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,8192,0.3174746566348606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,7168,0.022749332918061152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,6144,0.017665778597195942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,7168,0.012964444855848948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,7168,0.31536444028218585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,10240,0.024516445067193773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,6144,0.02202933364444309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,5120,0.015057777365048727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,6144,0.32789866129557294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,6144,0.01179377817445331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,5120,0.02048711147573259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,4096,0.011697777443461947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,5120,0.010336000058386061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,5120,0.32051287757025826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,4096,0.01891911029815674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,3584,0.01034577770365609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,4096,0.008931555681758458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,4096,0.3037475479973687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,3584,0.017674666312005784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,3072,0.009594666461149851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,3584,0.008661333057615492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,3584,0.33216622140672475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,3072,0.01648000048266517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,2560,0.008392888638708327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,3072,0.31900888019137913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,3072,0.008008889026112026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,2560,0.015448888142903646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,2048,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,2560,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,2560,0.32316711213853627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,2048,0.014120888378885059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,1536,0.006607111129495833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,2048,0.007280000381999546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,2048,0.3372302320268419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,1536,0.013203555511103736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,1024,0.004868444469239977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,1536,0.006699555450015598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,1536,0.3229822317759196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,1024,0.012435555458068848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,768,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,1024,0.00628355559375551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,1024,0.33118221494886607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,768,0.012476444244384766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,512,0.0037475555307335327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,768,0.006276444428496891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,768,0.3227288987901476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,512,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,256,0.003277333246337043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,512,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,512,0.31415022744072807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,256,0.01182666669289271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,128,0.003198222153716617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,256,0.00589155571328269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,256,0.32472621070014107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,128,0.011428444749779172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,64,0.0028071111689011254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,6144,128,0.005592888842026393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,6144,128,0.3059973451826307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,64,0.011356444822417365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,6144,32,0.011506666739781698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,65536,0.10269422001308864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,65536,0.07339911328421699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,65536,0.06935466660393609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,16384,0.03152444296412998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,65536,0.3434186776479085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,6144,32,0.002924444360865487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,16384,0.02958044409751892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,12288,0.025736000802781846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,16384,0.020765332712067496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,16384,0.3238186571333143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,12288,0.024636444118287828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,10240,0.021720889541837905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,12288,0.016930666234758165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,12288,0.3286622100406223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,10240,0.021889777647124395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,10240,0.01532444523440467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,8192,0.019306666321224637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,10240,0.329257779651218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,8192,0.02088622252146403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,7168,0.01700444519519806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,8192,0.33876978026496035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,7168,0.022704000274340313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,7168,0.0123857781291008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,7168,0.3342595630221897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,6144,0.01556799974706438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,6144,0.020304888486862183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,6144,0.3189084529876709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,6144,0.011424889167149862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,5120,0.013659555878904132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,5120,0.01923111081123352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,5120,0.32175644238789874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,5120,0.009906666974226633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,4096,0.011344888971911537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,4096,0.018068444397714403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,4096,0.32211555374993217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,3584,0.009045333498054082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,3584,0.01684533390733931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,8192,0.013850666582584381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,3584,0.3173137770758735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,3072,0.008772444393899705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,3584,0.008340444829728868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,3072,0.016139555308553908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,2560,0.007667555577225155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,3072,0.007687110867765214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,2560,0.015092444088723926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,4096,0.008438222110271454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,3072,0.463114685482449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,2560,0.3246844344668918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,2048,0.007428444094128079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,2560,0.007311999797821045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,2048,0.013839999834696451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,2048,0.006960888703664144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,1536,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,2048,0.40479111671447754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,1536,0.012808889150619507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,1024,0.004559999952713649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,1536,0.33629867765638566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,1536,0.00664444433318244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,1024,0.012711111042234631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,768,0.004200889004601372
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,1024,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,1024,0.3249137666490343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,768,0.012395555774370829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,512,0.003744889050722122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,768,0.005991999887757831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,768,0.3476613362630208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,512,0.012072000238630505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,256,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,512,0.005930666708283954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,512,0.327078209982978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,256,0.011422221859296164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,128,0.0028844444702068963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,256,0.005734222216738596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,256,0.3506053288777669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,128,0.011390222443474663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,64,0.0029048888633648553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,5120,128,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,5120,128,0.3140604496002197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,5120,32,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,64,0.011364444262451597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,65536,0.08288088772031996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,5120,32,0.011815999945004782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,65536,0.06931733422809176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,16384,0.02576444380813175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,65536,0.06713777780532837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,16384,0.026390223039521113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,65536,0.34381333986918133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,16384,0.02032711108525594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,12288,0.021162667208247717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,12288,0.02269511090384589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,12288,0.34612711270650226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,12288,0.01643200053109063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,10240,0.017603556315104168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,10240,0.020656888683636982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,10240,0.014888889259762235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,10240,0.3324728806813558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,8192,0.01590844492117564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,16384,0.3345955477820502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,8192,0.019708444674809773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,7168,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,8192,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,8192,0.341398212644789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,7168,0.020998222960366145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,6144,0.012980444563759698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,7168,0.011643555429246692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,7168,0.3362737761603461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,6144,0.020045333438449435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,5120,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,6144,0.010977778169843884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,6144,0.2937564320034451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,5120,0.018872888551818002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,4096,0.012063110868136087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,5120,0.33564088079664445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,4096,0.01686755485004849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,4096,0.3109840022193061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,4096,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,3584,0.011349333657158745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,3584,0.015558222929636637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,3584,0.31474577056037056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,3584,0.007961778177155389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,5120,0.00965244405799442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,3072,0.014875556031862894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,3072,0.007639110916190677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,3072,0.30420178837246364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,2560,0.007269333634111617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,2560,0.013888888888888888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,2560,0.3338444497850206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,2560,0.007272000114123027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,2048,0.005994666781690385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,3072,0.010285333626800114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,2048,0.01349511080318027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,1536,0.00535377777285046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,2048,0.0069217777086628815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,2048,0.32710666126675075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,1536,0.012684444586435953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,1024,0.004231111043029361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,1536,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,1536,0.3272622161441379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,1024,0.012454222473833295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,768,0.003918222255176968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,1024,0.006230222268237009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,1024,0.32509512371487087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,768,0.012113778127564324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,512,0.003569777641031477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,768,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,768,0.33170755704243976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,512,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,256,0.003237333355678452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,512,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,256,0.01146133326821857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,512,0.46663464440239805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,256,0.005595555735958948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,128,0.0029315555261241067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,256,0.33418578571743435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,128,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,4096,128,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,64,0.0028728888266616394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,4096,32,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,4096,128,0.3908115492926703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,64,0.011047110789351992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,4096,32,0.011724444727102915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,65536,0.06698489189147949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,65536,0.06621422370274861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,16384,0.022971555590629578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,65536,0.3158293300204807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,16384,0.025412445267041523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,16384,0.32396088706122506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,12288,0.018416888184017606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,16384,0.01977599991692437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,65536,0.07227466503779094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,12288,0.0225262227985594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,12288,0.016175111134847004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,10240,0.016255110502243042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,12288,0.33422221077813047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,10240,0.020265777905782063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,8192,0.01423288881778717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,10240,0.014754666222466363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,10240,0.32821777131822377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,8192,0.019634667370054457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,7168,0.013631111217869652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,8192,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,8192,0.33787910143534344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,7168,0.020620443754725985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,7168,0.01182044463025199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,6144,0.012483555409643384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,7168,0.3354586760203044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,6144,0.018927110566033255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,5120,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,6144,0.01035288886891471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,6144,0.33675111664666074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,5120,0.017234666479958426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,4096,0.01205244412024816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,5120,0.009447111023796929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,5120,0.33053066995408803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,4096,0.015853333804342482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,3584,0.010803555448849997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,4096,0.008332444561852349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,4096,0.3335422144995795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,3584,0.014879999889267815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,3072,0.009720000127951304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,3584,0.007972444097201029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,3584,0.3279342121548123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,3072,0.014149333039919535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,2560,0.0069217777086628815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,3072,0.007617777420414819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,3072,0.32511644893222386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,2560,0.01365066650840971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,2560,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,2048,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,2560,0.32398756345113117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,2048,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,1536,0.004546666724814309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,2048,0.006968888971540663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,2048,0.2912151018778483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,1536,0.012950222525331708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,1536,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,1024,0.0042071110672420925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,1536,0.322687996758355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,1024,0.012149333126015134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,768,0.0038328886859946777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,1024,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,1024,0.3249982198079427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,768,0.012114666402339935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,512,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,768,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,768,0.32264799541897243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,512,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,256,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,512,0.0058604445722368025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,512,0.3195964495340983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,256,0.011722666521867117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,128,0.003165333428316646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,256,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,256,0.319598224427965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,128,0.011392888923486074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,64,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3584,128,0.005679111099905438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3584,32,0.0029066666546795103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3584,128,0.2916399902767605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,65536,0.06021333403057522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,64,0.011078221930397881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3584,32,0.01146311147345437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,65536,0.06515022118886311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,16384,0.01993777851263682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,65536,0.06515910890367296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,16384,0.02556444373395708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,65536,0.3330720000796848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,16384,0.019693333241674636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,12288,0.01722044414944119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,16384,0.3211679988437229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,12288,0.02204888893498315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,12288,0.01588444411754608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,10240,0.014836443795098199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,10240,0.0203155560625924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,12288,0.4734791119893392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,8192,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,10240,0.014045332868893942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,8192,0.01889866590499878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,8192,0.3327377902136909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,7168,0.01239999963177575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,8192,0.011534222298198275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,7168,0.018901333212852478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,7168,0.011181332998805575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,6144,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,10240,0.34196088049146867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,6144,0.017952000101407368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,6144,0.009697777529557547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,6144,0.3330666753980848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,5120,0.01036533299419615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,5120,0.016897777716318767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,5120,0.33817601203918457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,4096,0.01036444471942054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,5120,0.009533333281675974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,7168,0.38257779015435117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,4096,0.014648000399271647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,3584,0.009684444301658207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,4096,0.008082666330867344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,3584,0.014719110396173267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,3584,0.007976888782448238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,3072,0.008351999852392409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,3584,0.36452799373202854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,3072,0.013774221969975365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,3072,0.33931732177734375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,3072,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,2560,0.006694222076071634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,2560,0.013495999905798169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,4096,0.3341040081448025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,2048,0.005566222386227713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,2560,0.32959643999735516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,2560,0.007366221812036302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,2048,0.01310400002532535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,1536,0.004937777916590373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,2048,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,2048,0.3248426649305555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,1536,0.012757333616415659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,1024,0.003923555629120933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,1536,0.3275724516974555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,1024,0.012044444680213928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,1024,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,1024,0.3132924503750271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,768,0.003588444242874781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,768,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,768,0.3388960096571181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,768,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,512,0.003220444545149803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,512,0.011966221862369113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,1536,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,256,0.003272888975010978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,512,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,256,0.010772444307804108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,256,0.32997510168287486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,256,0.005610666755172942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,128,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,128,0.01076533314254549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,128,0.3022426764170329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,3072,128,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,64,0.01036266651418474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,3072,512,0.32898489634195965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,32,0.0028044444819291434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,65536,0.054126222928365074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,3072,32,0.011779555843936073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,65536,0.06246044238408407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,65536,0.06377333402633667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,65536,0.3323440021938748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,16384,0.01829688913292355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,3072,64,0.002542222125662698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,12288,0.015876443849669564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,16384,0.024028443627887305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,12288,0.021352888809310064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,12288,0.2855137719048394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,16384,0.4810933536953396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,10240,0.013086222112178802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,12288,0.015171556009186639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,10240,0.013454222016864352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,10240,0.01975288821591271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,8192,0.011956444217099084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,10240,0.313972446653578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,8192,0.018585777944988675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,8192,0.011196444431940714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,7168,0.011325333681371478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,8192,0.33456354671054417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,16384,0.01941333380010393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,7168,0.018252443936136033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,7168,0.010837333069907295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,7168,0.37688178486294216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,6144,0.016717332932684157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,6144,0.33516621589660645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,6144,0.009710222482681274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,5120,0.009729777773221334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,5120,0.015793777174419828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,5120,0.3298257721794976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,5120,0.009036444127559662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,4096,0.009397333694828881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,4096,0.014195554786258273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,6144,0.011235555840863122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,3584,0.0085235552655326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,4096,0.3004115687476264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,4096,0.008296888735559251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,3584,0.013887999786270989
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,3072,0.007857778006129796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,3584,0.007725333174069722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,3584,0.30196621682908803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,3072,0.013447110851605734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,3072,0.007688889073001013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,3072,0.34053778648376465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,2560,0.013425778183672162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,2560,0.007302222152551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,2560,0.3274079958597819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,2048,0.005583111196756363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,2048,0.012835555606418185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,2048,0.32766845491197377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,2048,0.006907555378145642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,1536,0.004549333204825719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,2560,0.006231110956933763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,1536,0.012799999780125089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,1024,0.003937777959638172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,1536,0.32143643167283803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,1536,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,1024,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,768,0.003866666720973121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,1024,0.006254222244024277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,1024,0.32032444741990834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,768,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,512,0.003556444413132138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,768,0.005922666854328579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,768,0.3358942137824164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,512,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,256,0.0032497776879204642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,512,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,512,0.31836620966593426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,256,0.011781333221329583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,128,0.0028968888024489083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,256,0.005566222386227713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,256,0.32617423269483775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,128,0.011515555282433828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,64,0.0026488889836602737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2560,128,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2560,128,0.2973911232418484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2560,32,0.0028168888141711554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,64,0.011122666299343109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,65536,0.04418577750523885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2560,32,0.011759999725553723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,65536,0.0611777769194709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,16384,0.015838222371207345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,65536,0.05893777476416694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,65536,0.3291573259565565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,16384,0.02372088862790002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,12288,0.012895111408498553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,16384,0.01903377804491255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,16384,0.33138932122124565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,12288,0.020933333370420668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,10240,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,12288,0.014449778530332776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,12288,0.31824445724487305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,10240,0.01941333380010393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,8192,0.01111644423670239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,10240,0.3258559968736437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,10240,0.012898666991127862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,8192,0.01773244473669264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,7168,0.010412444670995077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,8192,0.010808889236715106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,8192,0.31625154283311635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,7168,0.016884444488419425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,6144,0.012177777787049612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,7168,0.01069866700304879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,7168,0.32247289021809894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,6144,0.015336889359686108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,5120,0.011364444262451597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,6144,0.009679999616410997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,6144,0.31563644938998753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,5120,0.014831110835075378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,4096,0.010419555836253695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,5120,0.009296889106432596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,5120,0.3252844545576307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,4096,0.014157333307796054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,3584,0.009329777624871995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,4096,0.008002666963471307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,4096,0.32090844048394096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,3584,0.013952889376216464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,3072,0.006656888872385025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,3584,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,3584,0.3215964370303684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,3072,0.01349511080318027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,2560,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,3072,0.007656888829337225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,3072,0.3280666669209798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,2560,0.01313511116637124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,2560,0.007194666398896112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,2560,0.3195253213246663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,2048,0.0052800000541739995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,2048,0.013092444174819522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,1536,0.004553777890072929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,2048,0.006912888752089606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,2048,0.33655110994974774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,1536,0.012432888978057437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,1024,0.004226666771703296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,1536,0.006603555546866522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,1024,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,1536,0.46512354744805234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,768,0.0038657776183552216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,1024,0.3350755638546414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,1024,0.006275555739800136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,768,0.012172444827026792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,768,0.0058986664646201665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,512,0.0035706667436493766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,768,0.32585867245992023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,512,0.011895111037625207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,256,0.003216888962520493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,512,0.005639111002286275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,512,0.3870897822909885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,256,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,128,0.0029351111087534162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,256,0.29376265737745494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,256,0.005624889085690181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,128,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,64,0.0026586666289303037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,2048,128,0.005609777652555042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,2048,32,0.0029111111329661477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,2048,128,0.2968204551272922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,64,0.011395555403497485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,65536,0.0340266658200158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,2048,32,0.01110666659143236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,65536,0.05992266866895887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,16384,0.01309333327743742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,65536,0.05820888943142361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,65536,0.3175457848442925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,16384,0.02335022224320306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,12288,0.015825778245925903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,16384,0.018280888597170513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,16384,0.3568444516923692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,12288,0.020676443974177044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,10240,0.014114666316244336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,12288,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,12288,0.337919102774726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,10240,0.018216000662909616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,8192,0.012727999852763282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,10240,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,10240,0.3654924498664008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,8192,0.01657511128319634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,7168,0.011376000112957425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,8192,0.3233964443206787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,8192,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,7168,0.01584533353646596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,6144,0.010510221951537661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,7168,0.010399999717871347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,7168,0.3260462284088135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,6144,0.015440889530711703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,5120,0.009542221824328104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,6144,0.009709333380063375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,6144,0.3393128977881537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,5120,0.014854222536087036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,4096,0.00870488915178511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,5120,0.009371555513805812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,5120,0.33239732848273384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,4096,0.014104889498816596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,3584,0.0069439998931354946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,4096,0.008279110822412703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,4096,0.33226490020751953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,3584,0.01385600037044949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,3072,0.006591111007663939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,3584,0.008007999923494127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,3584,0.3270435598161486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,3072,0.013445333474212222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,2560,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,3072,0.0074942219588491656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,3072,0.3334346612294515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,2560,0.013096888860066732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,2048,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,2560,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,2560,0.323853333791097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,2048,0.012911111116409302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,1536,0.004547555413511065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,2048,0.0069368887278768755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,2048,0.32828089925977916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,1536,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,1024,0.0039048890272776284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,1536,0.006773333168692059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,1536,0.33128266864352757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,1024,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,768,0.0035911111368073356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,1024,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,1024,0.3282613224453396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,768,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,512,0.0035235554807715942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,768,0.31841156217787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,512,0.011720889144473605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,512,0.3317760096655952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,512,0.005879111174080107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,768,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,256,0.0031191110610961914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,256,0.011757333245542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,256,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,128,0.0028488888508743713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,128,0.011386666860845355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,256,0.47672176361083984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1536,128,0.2931431134541829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1536,128,0.005584888988071018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,32,0.002883555574549569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,64,0.011462222370836469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,65536,0.025065778030289546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1536,32,0.011144888897736868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,65536,0.05686133437686496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,65536,0.34603201018439395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,16384,0.014723555909262763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,65536,0.05789599816004435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,16384,0.022335110439194575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1536,64,0.0029128889242808023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,12288,0.01182666669289271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,16384,0.3331129021114773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,16384,0.016525333126386006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,12288,0.018844443890783522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,12288,0.013734222286277346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,10240,0.011085333095656501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,12288,0.32472888628641766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,10240,0.017557332913080852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,10240,0.012404444317022959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,8192,0.010416888528399997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,10240,0.39066754447089297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,8192,0.01651733285850949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,7168,0.009505777723259395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,8192,0.33998576800028485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,8192,0.010737777584128909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,7168,0.015840889679061044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,6144,0.0086666668454806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,7168,0.010091555615266165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,7168,0.3275573253631592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,6144,0.015284443895022074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,5120,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,6144,0.009701333112186855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,6144,0.3490026791890462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,5120,0.01479288935661316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,4096,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,5120,0.008962666822804345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,5120,0.32896799511379665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,4096,0.014122666584120857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,3584,0.006928000185224745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,4096,0.008386666576067606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,4096,0.34361865785386825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,3584,0.013761777844693927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,3072,0.006592888798978593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,3584,0.007727999654081132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,3584,0.3436720106336806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,3072,0.013448889056841532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,2560,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,3072,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,3072,0.33610044585333926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,2560,0.013089777694808112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,2048,0.00665511108107037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,2560,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,2560,0.32780800925360787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,2048,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,1536,0.005657777604129579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,2048,0.3224284383985731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,1536,0.012127111355463663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,1536,0.32516177495320636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,1024,0.004648000001907349
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,1024,0.012173333101802401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,2048,0.01276000009642707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,1024,0.3407271173265245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,1024,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,768,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,768,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,768,0.3276133272382948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,768,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,512,0.003607999947335985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,1536,0.00666933341158761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,512,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,512,0.005597333527273602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,512,0.3228542274898953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,256,0.011373333632946014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,128,0.0031519999934567344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,256,0.31109510527716744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,128,0.011372444530328115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,128,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,64,0.0029893333299292457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,1024,128,0.29431109958224827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,32,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,64,0.011100444528791638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,65536,0.03156177865134345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,1024,32,0.010688888529936472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,1024,256,0.0032017777363459268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,65536,0.05542044507132637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,65536,0.05755822526084053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,16384,0.012484444512261285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,65536,0.32396088706122506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,16384,0.02198311189810435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,16384,0.3420737849341498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,12288,0.010749333434634738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,12288,0.018606222338146634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,12288,0.3293244573805067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,12288,0.013750221994188098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,1024,256,0.005583111196756363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,10240,0.009329777624871995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,16384,0.01649777756796943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,10240,0.017636444833543565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,8192,0.009296000003814697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,10240,0.33571910858154297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,10240,0.012471111284361945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,8192,0.016202666693263583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,7168,0.00868711123863856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,8192,0.011077333655622272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,8192,0.32984089851379395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,7168,0.015783111254374187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,6144,0.008007999923494127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,7168,0.010354667074150508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,7168,0.3420257833268907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,6144,0.01521511044767168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,6144,0.009677333136399588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,5120,0.007608888877762689
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,6144,0.3376035425398085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,5120,0.014800000521871777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,4096,0.007643555601437886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,5120,0.009404444032245213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,5120,0.3540364371405707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,4096,0.014119999276267158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,3584,0.007159111400445302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,4096,0.008302222523424361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,4096,0.3356000052558051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,3584,0.013796444568369122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,3072,0.0069857777820693115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,3584,0.007939555578761632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,3584,0.3433546755048964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,3072,0.013455111119482251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,2560,0.006663110935025745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,3072,0.007497777541478475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,3072,0.336641788482666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,2560,0.013073777986897362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,2048,0.0052328887912962176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,2560,0.007319110963079665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,2560,0.33300712373521596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,2048,0.01273155543539259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,1536,0.0052391112678580815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,2048,0.006931555353932911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,2048,0.334775103463067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,1536,0.012450666891203986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,1024,0.004270222038030624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,1536,0.006625777731339137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,1536,0.33203911781311035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,1024,0.012364444633324942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,768,0.003882666842805015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,1024,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,1024,0.3310515615675184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,768,0.012102222277058495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,512,0.003540444291300244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,768,0.0058986664646201665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,768,0.3346871005164252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,512,0.011720000041855706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,256,0.0032248888164758682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,512,0.005624889085690181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,512,0.3273288938734266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,256,0.01146577795346578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,128,0.002992000016901228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,256,0.005565333283609814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,256,0.33092355728149414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,128,0.011445333560307821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,64,0.002847111059559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,768,128,0.30569156010945636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,768,32,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,64,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,65536,0.023362666368484497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,768,32,0.01108888867828581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,65536,0.05491200089454651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,65536,0.05413777629534403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,65536,0.33711732758416074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,16384,0.010739555789364709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,16384,0.021333333518770006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,16384,0.32683467864990234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,768,128,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,12288,0.008752889103359645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,12288,0.018605333235528734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,12288,0.29541511005825466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,10240,0.008013333711359236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,10240,0.017633777525689866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,10240,0.3147359953986274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,16384,0.016207110550668504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,8192,0.008390222158696916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,10240,0.012455110748608908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,8192,0.01071200023094813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,8192,0.016190222567982145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,7168,0.008438222110271454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,8192,0.3184000121222602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,7168,0.015568888849682279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,7168,0.01032977799574534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,6144,0.007612444460391998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,7168,0.31933334138658315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,6144,0.015184889237085978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,6144,0.009679111341635386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,5120,0.007166222565703922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,5120,0.01479466590616438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,12288,0.01348622226052814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,5120,0.2936951054467095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,5120,0.00902844468752543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,4096,0.007337777978844113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,4096,0.014109333356221518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,4096,0.312570677863227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,3584,0.006980444822046492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,4096,0.008263111114501953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,6144,0.4050515492757161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,3584,0.013415999710559845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,3072,0.006598222172922558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,3584,0.007671111159854465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,3584,0.34352888001335996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,3072,0.013463999662134381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,2560,0.00665511108107037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,3072,0.007619555625650618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,3072,0.37050310770670575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,2560,0.012978666358523898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,2048,0.00601244428091579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,2560,0.0069466663731469055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,2560,0.34075289302402073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,2048,0.012835555606418185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,1536,0.004579555657174853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,2048,0.006656000183688269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,2048,0.3502515686882867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,1536,0.012140444583363004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,1024,0.003909333298603694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,1536,0.006272000157170826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,1536,0.3370773262447781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,1024,0.012105777859687805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,768,0.003555555724435382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,1024,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,1024,0.3395804564158122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,768,0.011714666253990598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,512,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,768,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,768,0.3362862269083659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,512,0.01146755533085929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,256,0.0032106666929192017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,512,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,512,0.3332737816704644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,256,0.0052684446175893145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,256,0.32830490006340873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,128,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,128,0.011419555379284753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,512,128,0.3071688810984294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,512,128,0.005630222045713001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,64,0.0025111111915773815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,512,32,0.002847111059559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,64,0.011042666931947073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,32,0.01071466671095954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,512,256,0.011350221931934357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,65536,0.05275377631187439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,65536,0.04934222168392605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,65536,0.3158257802327474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,16384,0.008749333520730337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,16384,0.020992888344658747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,16384,0.3396000067392985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,16384,0.01624000072479248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,12288,0.00831200016869439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,65536,0.01519377695189582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,12288,0.018588445252842374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,10240,0.007622222105662028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,12288,0.013493333425786761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,12288,0.3322266737620036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,10240,0.01741155485312144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,8192,0.007266666326257918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,10240,0.3367662164900038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,8192,0.01620533400111728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,8192,0.010774222513039907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,8192,0.32303910785251194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,7168,0.015625778171751235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,7168,0.3034755653805203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,10240,0.012065778175989786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,6144,0.006280000011126201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,7168,0.010321777727868822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,6144,0.014871110518773397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,7168,0.006650666809744305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,6144,0.00944088896115621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,6144,0.33288711971706814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,5120,0.006304889089531369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,5120,0.014476444986131458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,5120,0.009029332962301042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,4096,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,5120,0.32726311683654785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,4096,0.013833777772055732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,3584,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,4096,0.008216000265545314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,4096,0.3330746756659614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,3072,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,3584,0.01348533315791024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,3072,0.01312444441848331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,3584,0.46889686584472656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,3072,0.3339075512356228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,2560,0.006339555399285422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,2560,0.012863111164834766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,2560,0.4213875664605035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,2560,0.007012444237867991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,2048,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,2048,0.012896888785892062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,3072,0.007339555356237624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,2048,0.006727111008432176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,1536,0.0045422220395671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,2048,0.36918311648898655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,1536,0.012126222252845764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,1536,0.006253333141406377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,3584,0.007699555820888943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,1536,0.35630755954318577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,1024,0.00388711111413108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,1024,0.012039110892348819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,1024,0.006019555446174409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,768,0.0035146665241983202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,1024,0.3408533202277289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,768,0.011378666592968835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,512,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,768,0.005980444451173146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,768,0.32636621263292104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,512,0.01110400011142095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,256,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,512,0.005675555517276128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,512,0.35062487920125324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,256,0.010421333213647207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,128,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,256,0.005581333405441708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,128,0.010381333529949188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,256,128,0.005233777893914117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,128,0.31558044751485187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,64,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,256,32,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,64,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,65536,0.014167999227841696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,256,32,0.01035733355416192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,65536,0.05070933368470934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,65536,0.3377262221442328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,256,256,0.3279377884334988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,16384,0.007701333198282454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,65536,0.04902311166127523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,16384,0.016532444291644625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,12288,0.007649777664078607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,16384,0.317392004860772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,12288,0.018977777825461496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,12288,0.323907560772366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,12288,0.013529777526855469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,10240,0.0070168889231152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,10240,0.01794933279355367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,10240,0.3181146780649821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,8192,0.006654222392373615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,10240,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,16384,0.020823111136754353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,8192,0.015844444433848064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,7168,0.006281777802440856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,8192,0.0107342220014996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,8192,0.3389297856224908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,7168,0.015568888849682279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,6144,0.006240889016124938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,7168,0.01035288886891471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,7168,0.3429697884453668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,6144,0.014800000521871777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,5120,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,6144,0.3302328851487902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,5120,0.014479110638300577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,5120,0.00870666652917862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,5120,0.33419733577304417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,4096,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,4096,0.014214222629865011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,4096,0.32467733489142525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,3584,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,3584,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,6144,0.009678222239017487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,3584,0.00775999989774492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,3072,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,3584,0.32246579064263237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,3072,0.01313333296113544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,3072,0.3331662284003364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,3072,0.007575111256705389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,4096,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,2560,0.01313511116637124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,2560,0.007369777394665613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,2560,0.3362746768527561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,2048,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,2048,0.012792888614866467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,1536,0.004561777744028303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,2048,0.0070257774657673305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,2048,0.3397031095292833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,1536,0.01276711126168569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,1024,0.003867555409669876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,1536,0.006481777876615524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,2560,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,1024,0.012094222009181976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,768,0.003489777859714296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,1536,0.44013065761990017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,1024,0.00590133335855272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,1024,0.33140267266167533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,512,0.0032702222880389956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,768,0.005616000129116907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,768,0.012119111087587146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,512,0.011678222152921887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,768,0.4049626721276178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,256,0.002954666606254048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,512,0.3310933378007677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,512,0.005692444327804778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,256,0.011287111375066968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,128,0.002954666606254048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,256,0.32205867767333984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,128,0.01146399974822998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,128,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,64,0.002569777684079276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,64,0.011011555790901184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,128,32,0.0025680000997251934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,128,32,0.011354666617181567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,16,128,256,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,65536,0.04963111215167575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,16384,0.007294221884674496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,16,128,128,0.2980497678120931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,12288,0.006279110908508301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,16384,0.02092622220516205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,10240,0.006809777683681912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,12288,0.01824355622132619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,8192,0.00628355559375551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,10240,0.017582222819328308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,7168,0.006317333214812809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,8192,0.015849777393870883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,65536,0.010427555276287926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,6144,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,7168,0.015569777952300178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,5120,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,4096,0.005576000031497743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,6144,0.01514400045077006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,5120,0.014560000763999091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,3584,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,3072,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,4096,0.01383733335468504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,2560,0.005680888891220093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,3584,0.013717333475748697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,2048,0.005731555736727185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,3072,0.013388444152143268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,1536,0.004625777817434735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,2560,0.012825777961148156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,1024,0.003859555555714501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,2048,0.012804444465372296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,768,0.003564444267087512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,1536,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,512,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,1024,0.011496889094511667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,256,0.0029093333416514923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,768,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,128,0.0029271110478374693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,512,0.01145333300034205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,64,0.0025271111064487034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,256,0.0103511114915212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,64,32,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,128,0.011429333024554782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,64,0.01108088923825158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,64,32,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,16384,0.0062871111763848206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,65536,0.049636443456014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,12288,0.006261333409282897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,16384,0.020672889219390023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,10240,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,12288,0.018581334087583754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,8192,0.005595555735958948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,10240,0.01690933273898231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,7168,0.005607999861240387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,8192,0.016237333416938782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,6144,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,7168,0.015573332707087198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,5120,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,6144,0.01513777838812934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,65536,0.009660444325870937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,4096,0.005968889014588461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,3584,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,5120,0.014472888575659858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,3072,0.005595555735958948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,4096,0.013838222457302941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,2560,0.006264888991912206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,3584,0.013457777599493662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,2048,0.00563733321097162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,3072,0.013154666456911298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,1536,0.0042257776690853965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,2560,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,1024,0.003919110943873723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,2048,0.012714666624863943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,768,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,1536,0.012145777543385824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,512,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,1024,0.012067555553383298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,256,0.0031644445326593188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,768,0.011759999725553723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,128,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,512,0.011247999966144562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,64,0.0024951110697454875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,256,0.011416888899273343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,16,32,32,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,128,0.011396444506115384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,64,0.01035288886891471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,16,32,32,0.011699555648697747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,16384,0.3104320102267795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,16384,0.16009422143300375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,16384,0.14299466874864367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,12288,0.24007911152309844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,12288,0.1907644404305352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,16384,0.34949154324001735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,12288,0.10879288779364692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,10240,0.2033768892288208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,10240,0.13783289326561823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,12288,0.33409423298305935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,8192,0.14614221784803602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,10240,0.0918053322368198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,8192,0.12168089548746745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,10240,0.4720471170213487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,8192,0.07568444146050347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,7168,0.12764533360799155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,8192,0.34169599745008683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,7168,0.13598755995432535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,6144,0.11099021964603001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,7168,0.06599111027187772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,6144,0.10157066583633423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,7168,0.41287822193569607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,5120,0.09251199828253852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,6144,0.057986663447486035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,5120,0.08960800038443671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,6144,0.3384275436401367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,4096,0.0786000026596917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,5120,0.04981866810056898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,5120,0.32803556654188365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,4096,0.07576088772879706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,3584,0.06641955508126153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,4096,0.04052622119585673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,3584,0.07138666841718885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,4096,0.3618426587846544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,3072,0.05765866571002536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,3584,0.03605066736539205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,3584,0.32709866099887425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,3072,0.05593333641688029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,2560,0.04862311151292589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,3072,0.03232266505559286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,3072,0.3355262279510498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,2560,0.04774844315316942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,2048,0.043856001562542386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,2560,0.02757866680622101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,2048,0.041379554404152766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,2560,0.32006933954026967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,1536,0.030930668115615845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,2048,0.029234667619069416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,2048,0.3300204541948107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,1536,0.04004977809058295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,1024,0.02208799951606327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,1536,0.019170666734377544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,1536,0.32988890012105304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,1024,0.023755555351575215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,768,0.018159111340840656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,1024,0.018153778380817838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,768,0.02437688906987508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,512,0.013029333617952136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,768,0.01535022258758545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,1024,0.4527164565192328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,768,0.33114756478203666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,256,0.008353778057628209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,512,0.01904977858066559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,512,0.013008000122176277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,512,0.29874666531880695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,256,0.01644088824590047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,128,0.006216889040337668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,256,0.010114666488435533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,256,0.3255511124928792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,128,0.014132445057233175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,64,0.005571555760171678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,65536,32,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,65536,128,0.2896648777855767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,64,0.01593155496650272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,65536,32,0.014518222875065274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,65536,0.3192613389756945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,65536,0.16491377353668213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,16384,0.08718044228023952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,16384,0.05428444345792135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,65536,0.16661511527167425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,65536,128,0.010196444061067369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,16384,0.05491022268931071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,12288,0.06763822502560086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,16384,0.33195556534661186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,12288,0.0628417796558804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,10240,0.058323555522494845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,12288,0.03586311141649882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,65536,0.33010933134290904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,12288,0.38431111971537274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,10240,0.05345333284801907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,8192,0.049569778972201876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,10240,0.033831109603246055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,10240,0.32408534155951607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,7168,0.03562755717171563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,8192,0.027905778752432928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,8192,0.36071377330356175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,7168,0.046625776423348315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,6144,0.03154755632082621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,7168,0.022738666998015508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,7168,0.3273235427008735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,6144,0.04162311222818162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,6144,0.3189351028866238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,6144,0.02068177858988444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,5120,0.02664622167746226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,8192,0.05134133497873942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,5120,0.036208887894948326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,4096,0.022648889157507155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,5120,0.01846933364868164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,5120,0.3148889011806912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,4096,0.028353777196672227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,3584,0.01998844411638048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,4096,0.015584889385435315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,4096,0.32380355728997123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,3584,0.028190222051408555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,3072,0.01741333305835724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,3584,0.014506666196717156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,3584,0.33119111590915257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,3072,0.023732443650563557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,2560,0.015463999576038785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,3072,0.013245333400037555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,3072,0.3253600067562527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,2560,0.021399999658266704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,2048,0.01570044457912445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,2560,0.3257404433356391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,2048,0.01991555591424306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,2048,0.012849777936935425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,2048,0.324139568540785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,1536,0.010710222025712332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,1536,0.01866755551762051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,1536,0.34743467966715497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,1536,0.010737777584128909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,1024,0.007851555943489075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,2560,0.012168889244397482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,1024,0.01455200049612257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,768,0.0069226668112807805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,1024,0.33008265495300293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,768,0.014814222852389017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,768,0.008036444584528605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,768,0.32133600446912974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,512,0.004577777865860197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,512,0.014081777797804939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,512,0.3181626796722412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,512,0.007278222176763747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,256,0.003536888708670934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,256,0.012484444512261285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,1024,0.009003555609120263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,256,0.32255199220445424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,128,0.0032248888164758682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,128,0.011716444459226398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,16384,128,0.3196293248070611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,128,0.006739555547634761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,64,0.0030586665703190696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,16384,32,0.0028728888266616394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,64,0.01273155543539259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,16384,32,0.011409777734014722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,65536,0.2282302114698622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,65536,0.14644711547427708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,16384,256,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,16384,0.06526399983300103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,16384,0.048356443643569946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,65536,0.14377600616878933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,16384,0.045964443021350436
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,12288,0.05186400148603651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,16384,0.32588354746500653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,12288,0.05231644378768074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,65536,0.34138933817545575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,12288,0.03126666612095303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,10240,0.04362755682733324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,12288,0.3114711178673638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,10240,0.04636266827583313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,10240,0.026465778549512226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,10240,0.3049982123904758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,8192,0.03932888971434699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,8192,0.022827555735905964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,8192,0.3113244374593099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,7168,0.03201155530081855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,7168,0.0395671096112993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,7168,0.3243742254045275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,7168,0.02033955521053738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,6144,0.02870044443342421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,8192,0.036722666687435575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,6144,0.03384533193376329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,5120,0.0243520008193122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,6144,0.018575110369258456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,6144,0.3251155482398139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,4096,0.021682666407691106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,5120,0.028209778997633193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,5120,0.31787289513481987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,4096,0.026214222113291424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,3584,0.018904000520706177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,4096,0.014214222629865011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,3584,0.02555466691652934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,3584,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,3584,0.3267529010772705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,3072,0.01702311138312022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,3072,0.02144711050722334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,3072,0.3263813389672173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,3072,0.012346666720178394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,2560,0.014961777461899651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,5120,0.016710221767425537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,2560,0.020921778347757127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,4096,0.3352675437927246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,2560,0.32464800940619576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,2560,0.01108088923825158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,2048,0.01827911039193471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,2048,0.32780443297492134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,2048,0.01111644423670239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,1536,0.008344000412358178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,1536,0.01756977703836229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,1536,0.3267742262946235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,1536,0.009719111025333405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,1024,0.006656888872385025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,1024,0.013895110951529609
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,2048,0.011239111423492432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,768,0.005960000058015187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,1024,0.32462043232387966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,1024,0.00833511104186376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,768,0.014153778553009033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,768,0.007667555577225155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,768,0.31904533174302846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,512,0.013445333474212222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,512,0.32356354925367564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,256,0.0032044444233179092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,256,0.012072000238630505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,256,0.3223182095421685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,512,0.004120888809363048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,128,0.003003555453485913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,256,0.006649777707126405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,128,0.011414222419261932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,512,0.006963555183675554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,12288,128,0.006495999793211619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,64,0.002858666703104973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,12288,128,0.30685777134365505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,12288,32,0.00282577777074443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,64,0.012431999875439538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,12288,32,0.011095999843544431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,65536,0.19143822458055285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,65536,0.1402319934633043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,16384,0.05685333410898844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,65536,0.14705599678887263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,16384,0.046810666720072426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,16384,0.038463112380769514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,12288,0.04287822379006279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,16384,0.3297146691216363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,65536,0.32855378256903756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,12288,0.04618933465745714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,10240,0.03702933258480496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,12288,0.02954933378431532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,12288,0.3126595550113254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,10240,0.04152177770932516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,8192,0.03040800160831875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,10240,0.02583911187118954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,10240,0.3254506587982178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,8192,0.03502755694919162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,7168,0.027875555886162653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,8192,0.021967111362351313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,8192,0.3186675442589654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,7168,0.031231999397277832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,6144,0.025832000705930922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,7168,0.019817777805858187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,7168,0.33656978607177734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,6144,0.027320888307359483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,5120,0.021771555145581562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,6144,0.017855111095640395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,6144,0.3157324526045057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,5120,0.02444000045458476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,5120,0.015993777248594496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,4096,0.018901333212852478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,5120,0.3297635449303521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,4096,0.023355555203225877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,3584,0.017197334104114108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,4096,0.013738666971524557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,3584,0.022096888886557683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,3584,0.3264133400387234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,3584,0.012915555801656512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,3072,0.015227556228637695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,3072,0.019966221517986722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,3072,0.32113422287835014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,3072,0.012723555167516073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,2560,0.012111110819710625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,2560,0.018575999471876357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,4096,0.29866133795844185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,2048,0.011462222370836469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,2560,0.3215866618686252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,2560,0.011602666642930774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,1536,0.008598222500748104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,2048,0.01684888866212633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,2048,0.01020088874631458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,2048,0.3088657855987549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,1024,0.0069724445541699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,1536,0.3254062334696452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,1536,0.009156444834338294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,1024,0.013703111145231458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,768,0.005803555664088991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,1024,0.007696000238259633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,768,0.013153778182135688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,1024,0.3171573215060764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,512,0.004600888739029567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,768,0.007295110987292395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,512,0.012736000120639801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,768,0.41587199105156797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,1536,0.016283555163277518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,512,0.006607999818192587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,256,0.011446221835083432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,256,0.3570275571611192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,256,0.006299555715587404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,128,0.003157333367400699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,512,0.34040533171759707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,256,0.0031679999083280563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,128,0.01071466671095954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,64,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,10240,128,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,10240,32,0.0028533333291610083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,10240,128,0.3088657855987549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,64,0.011064888702498542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,10240,32,0.011459555890825061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,65536,0.15541066063774958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,65536,0.09070399734708999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,16384,0.04561422268549601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,65536,0.07801422145631579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,16384,0.03391822179158529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,16384,0.02459555533197191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,16384,0.3301431073082818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,12288,0.03558399942186143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,65536,0.3359333409203424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,12288,0.03698755635155572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,10240,0.02716355522473653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,12288,0.3501146634419759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,10240,0.032292445500691734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,10240,0.016605334149466622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,8192,0.022338666849666174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,10240,0.3302728864881727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,8192,0.029300444655948218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,8192,0.32038132349650067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,8192,0.015307555596033732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,7168,0.020136000381575692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,12288,0.01872977779971229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,7168,0.024849777420361836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,6144,0.01824977828396691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,7168,0.013502221968438892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,7168,0.31535198953416615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,6144,0.023776888847351074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,5120,0.015199111567603217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,6144,0.01241422196229299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,6144,0.3339564535352919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,5120,0.02196088929971059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,4096,0.013823111024167804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,5120,0.01162755572133594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,5120,0.3248159885406494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,4096,0.02030311193731096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,3584,0.011926222178671094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,4096,0.009695111049546136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,4096,0.3281324439578586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,3584,0.018938667244381376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,3072,0.010640000303586325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,3584,0.009373333719041612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,3584,0.324619558122423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,3072,0.017321777012613084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,2560,0.009567110902733272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,3072,0.008665777742862701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,3072,0.32871998680962455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,2560,0.01684888866212633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,2048,0.00903200027015474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,2560,0.009015111459626092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,2560,0.3232906659444173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,2048,0.015221332510312399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,1536,0.006968888971540663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,2048,0.29791556464301217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,1536,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,1536,0.29477866490681964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,1536,0.0069013333155049225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,1024,0.005015111217896144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,1024,0.013076444466908773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,1024,0.3075644440121121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,768,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,768,0.012758221891191272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,2048,0.007652444144090016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,768,0.006281777802440856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,512,0.003539555602603488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,512,0.011732444167137146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,1024,0.006624889042642381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,512,0.0063528890411059065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,256,0.0031973332580592898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,256,0.011744000017642975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,256,0.3203929000430637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,256,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,128,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,768,0.3319759898715549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,128,0.011058666639857821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,64,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,8192,128,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,128,0.3021093209584554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,8192,32,0.002830222249031067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,64,0.012042666474978128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,65536,0.1377413272857666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,8192,32,0.010728889041476779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,8192,512,0.32396888732910156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,65536,0.08374489016003078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,16384,0.038112001286612615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,65536,0.07540533277723524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,16384,0.03204888767666287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,16384,0.2984000047047933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,16384,0.022707555029127333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,12288,0.030423998832702637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,65536,0.32388001018100315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,12288,0.03810844487614102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,10240,0.0269128895468182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,12288,0.29956176545884877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,12288,0.017864000466134813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,10240,0.03228977653715346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,8192,0.022384888596004907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,10240,0.0162471118900511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,10240,0.300144010119968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,8192,0.027179555760489568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,7168,0.020230222079488967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,8192,0.01443377799457974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,8192,0.32726044125027126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,7168,0.013168888787428537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,6144,0.016571554872724745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,7168,0.30771199862162274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,6144,0.023846222294701472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,6144,0.3135911093817817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,6144,0.012128888732857175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,5120,0.014107555150985718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,5120,0.021957332889238994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,5120,0.011092444260915121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,5120,0.32566311624315053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,4096,0.012823111481136747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,7168,0.026239110363854304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,4096,0.019724445210562814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,4096,0.009391999906963771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,3584,0.011164444188276926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,4096,0.32415999306572807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,3584,0.018212444252438016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,3584,0.008785777621799046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,3584,0.327078209982978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,3072,0.017142222987280954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,3072,0.00832444429397583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,2560,0.009143110778596666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,2560,0.016168889072206285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,2560,0.33054312070210773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,2560,0.00797155582242542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,2048,0.008014221986134848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,3072,0.3201013406117757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,2048,0.015238222148683337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,2048,0.007327111230956183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,1536,0.006628444625271692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,2048,0.31615644031100804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,1536,0.01385155568520228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,3072,0.010086221827401055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,1536,0.3132853243086073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,1024,0.0046826667255825466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,1536,0.006829333388143116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,1024,0.012455110748608908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,768,0.003908444609906939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,1024,0.3434782293107774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,768,0.012067555553383298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,768,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,768,0.3341964350806342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,512,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,512,0.012130666938092975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,512,0.33762134446038145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,256,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,1024,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,256,0.01074222226937612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,256,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,256,0.3229457802242703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,128,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,128,0.011553777588738335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,7168,128,0.30412265989515513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,64,0.0028577778074476454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,512,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,64,0.011104889214038849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,7168,32,0.002918222298224767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,65536,0.11978044774797227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,65536,0.07633955611122979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,7168,128,0.005904888941182031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,16384,0.035424000687069364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,16384,0.030792001220915053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,65536,0.07408622238371107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,65536,0.3375706672668457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,7168,32,0.009706666900051964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,12288,0.02871911062134637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,16384,0.022630222969584998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,12288,0.036334223217434354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,16384,0.3379048771328396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,10240,0.02402755618095398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,12288,0.01738755570517646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,12288,0.3239155610402425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,10240,0.028233776489893597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,10240,0.015867556134859722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,10240,0.2803448836008708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,8192,0.02511288887924618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,7168,0.018544000056054857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,8192,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,7168,0.025703110628657874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,8192,0.3819084432390001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,6144,0.017480888300471835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,7168,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,7168,0.30953865581088597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,6144,0.022083555658658344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,6144,0.3147733211517334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,5120,0.014686221877733866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,6144,0.011715555356608497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,5120,0.021287999219364587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,5120,0.36100178294711643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,5120,0.010627555350462595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,4096,0.011933333343929715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,4096,0.018900444110234577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,4096,0.3464284472995334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,4096,0.00903911143541336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,3584,0.009829333258999718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,3584,0.018434666925006442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,8192,0.020356444848908316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,3072,0.009471110999584198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,3584,0.29191554917229545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,3584,0.008658666577604082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,3072,0.01698666645420922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,2560,0.008667555948098501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,3072,0.008307555483447181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,3072,0.32064710723029244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,2560,0.015230221880806817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,2048,0.008021333151393467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,2560,0.007684444387753804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,2560,0.3276222281985813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,2048,0.01419911119672987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,2048,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,1536,0.006309333360857434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,2048,0.3261146545410156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,1536,0.01382844481203291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,1024,0.004586666822433472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,1536,0.006615110983451207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,1536,0.3333626588185628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,1024,0.012693333129088083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,768,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,1024,0.0063271112740039825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,1024,0.3258604473537869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,768,0.012415111064910889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,512,0.00350577798154619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,768,0.00600355573826366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,768,0.33402665456136066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,512,0.012103111379676394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,256,0.003257777748836411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,512,0.005973333285914526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,512,0.3231697877248128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,256,0.011739555332395764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,128,0.003072888900836309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,256,0.006001777946949005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,256,0.3301626576317681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,128,0.011379555695586734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,64,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,6144,128,0.005881777654091517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,6144,32,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,6144,128,0.3103120062086317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,64,0.012123555772834353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,65536,0.10080799791547988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,6144,32,0.011063999599880643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,65536,0.0714942216873169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,16384,0.03152533372243246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,65536,0.06940978103213839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,16384,0.029253333806991577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,16384,0.021196444829305012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,16384,0.33969510926140684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,12288,0.025183111429214478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,12288,0.03215288784768846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,12288,0.31072621875339085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,12288,0.01686311099264357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,10240,0.02887466549873352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,65536,0.3478631178538005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,10240,0.01516533394654592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,8192,0.019073777728610568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,10240,0.021576889687114294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,8192,0.02604444490538703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,8192,0.01351999988158544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,7168,0.016527111331621807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,8192,0.3277946578131782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,7168,0.02270488937695821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,10240,0.3274257712894016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,6144,0.015621332658661736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,7168,0.3220133251614041
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,7168,0.012252444194422828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,6144,0.021430222524536982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,6144,0.011286222272449069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,5120,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,6144,0.31159377098083496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,5120,0.019650666250122916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,4096,0.011132444772455426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,5120,0.01000088867213991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,5120,0.3356897830963135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,4096,0.018353778454992507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,3584,0.009617778162161509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,4096,0.008465777668688033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,4096,0.3148577743106418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,3584,0.017512000269360013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,3072,0.00850844466023975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,3584,0.00831911133395301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,3584,0.3283289008670383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,2560,0.007660444411966536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,3072,0.007642666498819987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,3072,0.0158933334880405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,3072,0.3293457825978597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,2560,0.014859555496109856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,2048,0.007647111184067196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,2560,0.007280888656775157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,2560,0.3241635428534614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,2048,0.01370844410525428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,1536,0.005595555735958948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,2048,0.006952889263629913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,2048,0.33031998740302193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,1536,0.013473778135246701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,1024,0.004306666553020477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,1536,0.3209795686933729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,1024,0.012455999851226807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,1024,0.006239999913507038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,1024,0.3214062319861518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,768,0.003879111260175705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,768,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,768,0.3212648762596978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,512,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,512,0.012059555285506778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,1536,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,512,0.005919999960396025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,512,0.3266791237725152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,256,0.0031742221779293488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,256,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,256,0.32396356264750165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,128,0.002918222298224767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,256,0.005927111125654644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,768,0.006004444426960415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,128,0.011358222199810876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,64,0.0028684443483750024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,5120,128,0.005592888842026393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,5120,32,0.0028195555011431375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,5120,128,0.3142533302307129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,64,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,65536,0.082168890370263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,5120,32,0.010679110884666443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,65536,0.06645688745710585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,16384,0.025410667061805725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,65536,0.0671973360909356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,16384,0.025992888543340895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,65536,0.3415164417690701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,16384,0.2767573462592231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,12288,0.020971554848882887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,16384,0.020279111133681405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,12288,0.028853333658642236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,12288,0.016364443633291457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,10240,0.017859554953045316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,12288,0.3248986668056912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,10240,0.02403555479314592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,8192,0.015855999456511605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,10240,0.014848000473446317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,10240,0.30973866250779897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,8192,0.02374222212367588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,8192,0.013080000049538083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,7168,0.014175999495718213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,8192,0.32161156336466473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,7168,0.02230577833122677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,6144,0.013387555877367655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,7168,0.011534222298198275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,7168,0.3307768768734402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,6144,0.020495999190542433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,5120,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,6144,0.010496888723638324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,6144,0.32634401321411133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,5120,0.0192577772670322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,5120,0.009184888667530483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,5120,0.33165955543518066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,4096,0.01715377800994449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,4096,0.3196817768944634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,3584,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,3584,0.016466667254765827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,3584,0.31985243161519367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,4096,0.01144177797767851
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,3072,0.010310221877362993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,3584,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,3072,0.014860444598727755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,4096,0.008039111064540016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,3072,0.007632888853549957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,2560,0.008818666968080733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,3072,0.3280675676133898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,2560,0.0143857780430052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,2560,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,2048,0.007879111501905654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,2560,0.309770663579305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,2048,0.013468444347381592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,1536,0.0064151109092765385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,2048,0.0068693334857622785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,2048,0.3278053336673313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,1536,0.012777778009573618
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,1024,0.005221333354711533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,1536,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,1536,0.31758756107754177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,1024,0.012267555627557965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,768,0.004335999902751711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,1024,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,1024,0.3146497673458523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,768,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,512,0.003912888881233004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,768,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,512,0.011789333489206104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,768,0.47812530729505753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,512,0.3228231006198459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,256,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,512,0.00564177789621883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,256,0.011558222273985544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,256,0.005618666609128316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,128,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,128,0.011064888702498542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,256,0.4176577726999919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,64,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,4096,128,0.2968977822197808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,4096,32,0.0028497777465316984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,64,0.010766222245163388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,4096,32,0.01054044481780794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,65536,0.07159822516971164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,65536,0.06435733371310763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,65536,0.06634400288263957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,16384,0.022077333596017625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,65536,0.37004711892869735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,4096,128,0.005594666633341048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,16384,0.3646124468909369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,16384,0.020244444410006206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,12288,0.01788444485929277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,12288,0.025435555312368605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,12288,0.3494257662031386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,10240,0.015828443898095023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,10240,0.025631111529138353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,16384,0.025461332665549383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,10240,0.014709333578745524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,10240,0.34191200468275285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,8192,0.01388177772363027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,8192,0.021360889077186584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,8192,0.32631201214260525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,12288,0.016345777445369296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,7168,0.013112000293201871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,8192,0.01239111108912362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,7168,0.020170667105250888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,7168,0.01070666644308302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,6144,0.012472000386979846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,7168,0.32009243965148926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,6144,0.0196115556690428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,5120,0.011490667031870948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,6144,0.01036977767944336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,6144,0.3482835557725694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,5120,0.018211555149820115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,4096,0.011015999648306104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,5120,0.009931555224789513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,5120,0.32979822158813477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,4096,0.015781333049138386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,3584,0.010696888797812991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,4096,0.008190222084522247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,4096,0.3362666765848796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,3584,0.015136000182893542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,3072,0.009344889058007134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,3584,0.007948444121413762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,3584,0.3281786706712511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,3072,0.014813333749771118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,2560,0.008339555727110969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,3072,0.007659555309348636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,3072,0.33486400710211855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,2560,0.014114666316244336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,2048,0.006736889067623351
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,2560,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,2560,0.3160586622026232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,2048,0.013091555900043912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,1536,0.005928888916969299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,2048,0.0069422221018208405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,2048,0.3305182192060682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,1536,0.0127688886390792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,1024,0.004884444591071871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,1536,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,1536,0.2878195444742839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,1024,0.01205599970287747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,1024,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,768,0.004518222063779831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,1024,0.33005865414937335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,768,0.012465777496496836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,512,0.0039013334446483185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,768,0.005951111101441913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,768,0.32649866739908856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,512,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,256,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,512,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,512,0.3257102171579997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,256,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,128,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,256,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,256,0.3256453408135308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,128,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,64,0.002565333412753211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3584,128,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3584,32,0.0028737777223189673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3584,128,0.29832445250617134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,64,0.011738667057620155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,65536,0.06201689110861885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3584,32,0.009737778041097853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,16384,0.020222221811612446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,65536,0.062295999791887074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,16384,0.02432800001568264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,65536,0.06504978073967828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,65536,0.3267137739393446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,16384,0.3218293454911974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,12288,0.01721333298418257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,16384,0.019692444139056735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,12288,0.025097777446111042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,12288,0.016152888536453247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,10240,0.014452444182501899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,12288,0.3189600043826633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,10240,0.023344889283180237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,10240,0.014446222119861178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,8192,0.012851555314328937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,10240,0.32661599583095974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,8192,0.020980444219377305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,8192,0.012127999630239276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,7168,0.012103111379676394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,8192,0.3863031069437663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,7168,0.019997333486874897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,6144,0.01237155579858356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,7168,0.01127022256453832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,7168,0.3369191222720676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,6144,0.018570666511853535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,5120,0.010057777994208867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,6144,0.010527999864684211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,6144,0.3539448844061957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,5120,0.016160888804329764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,4096,0.010428444378905825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,5120,0.34531378746032715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,5120,0.009331555830107795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,4096,0.0151235560576121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,3584,0.009364444348547194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,4096,0.00813066628244188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,4096,0.3368391195933024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,3584,0.01478488908873664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,3072,0.008301333420806462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,3584,0.008035555481910706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,3584,0.34561067157321507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,3072,0.013826666606797112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,2560,0.007282666862010956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,3072,0.0075599998235702515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,3072,0.3344559934404161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,2560,0.013637333280510373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,2048,0.006331555545330048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,2560,0.007241778075695038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,2560,0.3362773259480794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,2048,0.013159111142158508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,1536,0.006231110956933763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,2048,0.006638222270541721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,2048,0.33260801103379994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,1536,0.012822222378518848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,1024,0.005224888937340842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,1536,0.006614222294754452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,1024,0.012444444828563266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,1024,0.006264888991912206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,768,0.004469333423508538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,1024,0.3289297686682807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,768,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,768,0.31838756137424046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,768,0.006028444402747684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,512,0.0038968887594011095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,512,0.011778666741318174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,512,0.3239964379204644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,256,0.0034924443397257063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,1536,0.33690399593777126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,256,0.011408888631396823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,256,0.3073973390791151
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,256,0.005601777798599667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,128,0.0032142222755485107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,128,0.011039999624093374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,3072,128,0.28747110896640354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,64,0.0029200000895394217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,128,0.005626666463083691
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,3072,32,0.002806222273243798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,64,0.011730666789743634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,3072,512,0.005593777944644292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,3072,32,0.010315555665228102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,65536,0.05332533187336392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,65536,0.06085511048634847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,65536,0.0614373352792528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,16384,0.018900444110234577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,65536,0.34665777948167587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,16384,0.023754666248957317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,12288,0.015543111496501498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,16384,0.327310217751397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,16384,0.019521777828534443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,12288,0.02384800049993727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,10240,0.01349511080318027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,12288,0.01555999947918786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,12288,0.3227955500284831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,10240,0.02199733257293701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,8192,0.011904888682895236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,10240,0.013088888592190213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,10240,0.330077330271403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,8192,0.01958044370015462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,7168,0.011403555671374003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,8192,0.011408000356621213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,8192,0.33164623048570424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,7168,0.01923199991385142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,6144,0.01129066695769628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,7168,0.010671999719407825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,7168,0.3334648874070909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,6144,0.01662577854262458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,6144,0.00977600034740236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,6144,0.3309146563212077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,5120,0.00979466653532452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,5120,0.015497777197096082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,4096,0.010227555202113258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,5120,0.009135111338562435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,5120,0.33658933639526367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,4096,0.014779556128713818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,3584,0.008854221966531541
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,4096,0.008011555506123437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,3584,0.014472888575659858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,4096,0.4676524268256293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,3072,0.008341333104504479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,3584,0.007995555798212687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,3072,0.013839999834696451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,3072,0.3287279870775011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,2560,0.006969778074158563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,3072,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,2560,0.013432000246312885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,2560,0.007258666886223688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,2560,0.3811937703026666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,2048,0.006253333141406377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,3584,0.32613245646158856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,2048,0.012866666747464074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,1536,0.005624889085690181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,2048,0.006689777804745569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,1536,0.012472000386979846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,1536,0.3385546737247043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,1536,0.006607999818192587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,1024,0.004920000003443824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,1024,0.011756444142924415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,1024,0.3348417811923557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,1024,0.00628266649113761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,768,0.004269333349333869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,768,0.01148444414138794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,2048,0.3330266740587023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,512,0.004202666795916027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,768,0.3322426742977566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,768,0.005932444499598608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,512,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,256,0.003516444315512975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,512,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,512,0.32631201214260525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,256,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,128,0.0032177778581778207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,256,0.005631111148330901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,256,0.3344399929046631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,128,0.011086222198274402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,64,0.0026266665922270883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2560,128,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2560,128,0.3096817864312066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2560,32,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,64,0.011194667054547204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,65536,0.04414577616585625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2560,32,0.011052444577217102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,65536,0.05866222249137031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,16384,0.01579466627703773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,65536,0.059063110086652965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,65536,0.3367644415961371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,16384,0.02347022294998169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,12288,0.012727999852763282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,16384,0.018978666928079393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,16384,0.33751644028557676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,12288,0.021338666478792827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,10240,0.011378666592968835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,12288,0.015139554937680563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,12288,0.3274266719818115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,10240,0.019959999455346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,8192,0.011060444845093621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,10240,0.01240711079703437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,10240,0.3415413432651096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,8192,0.01818933255142636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,8192,0.011128889189826118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,8192,0.3330497741699219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,7168,0.017247110605239868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,7168,0.010384889112578498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,6144,0.010663111176755695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,7168,0.32217155562506783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,6144,0.015272888872358533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,6144,0.3225288920932346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,6144,0.009685333404276106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,5120,0.008731555607583787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,7168,0.010527999864684211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,5120,0.014829332629839579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,4096,0.010416000253624385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,5120,0.31496800316704643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,5120,0.009366222553782994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,4096,0.014171555638313293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,3584,0.009202666580677032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,4096,0.008286221987671321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,4096,0.30902134047614205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,3584,0.01348355578051673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,3072,0.006887110984987683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,3584,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,3584,0.34021690156724715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,3072,0.013468444347381592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,2560,0.005877333382765452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,3072,0.007709333466158972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,3072,0.33526934517754453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,2560,0.01311377767059538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,2560,0.007012444237867991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,2048,0.006697777658700943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,2560,0.317833767996894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,2048,0.012726222475369772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,1536,0.00565422202150027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,1536,0.012448000411192576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,2048,0.47562577989366317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,1536,0.006666666517655055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,1536,0.3327102131313748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,1024,0.012175111307038201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,2048,0.00693866651919153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,1024,0.3459786574045817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,768,0.004282666577233208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,1024,0.006239999913507038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,768,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,768,0.3168231116400825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,512,0.003552888830502828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,1024,0.004642666627963384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,512,0.01108088923825158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,512,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,512,0.38980889320373535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,256,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,256,0.011459555890825061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,256,0.31925421290927464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,256,0.005567111074924469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,128,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,768,0.005962666538026597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,128,0.011381333072980246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,64,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,2048,128,0.3026577896542019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,2048,128,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,2048,32,0.003076444483465619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,65536,0.033831109603246055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,64,0.011499555574523078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,2048,32,0.011038222246699862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,65536,0.05712088611390856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,16384,0.01294844432009591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,65536,0.05783200263977051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,65536,0.3273395432366265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,16384,0.02293866707219018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,16384,0.01722311145729489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,16384,0.331188440322876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,12288,0.021212443709373474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,12288,0.33065954844156903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,10240,0.013767110804716745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,10240,0.019462222854296368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,12288,0.011377777490350934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,10240,0.012287110918098025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,8192,0.01238044434123569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,12288,0.013682666752073499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,8192,0.01620444489849938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,8192,0.32435909907023114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,7168,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,7168,0.016203555795881484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,10240,0.3340008788638645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,7168,0.010355555348926121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,6144,0.01074488874938753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,7168,0.3198168807559543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,6144,0.015400000744395785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,6144,0.009675555759006077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,6144,0.33248265584309894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,5120,0.009391110804345872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,8192,0.010710222025712332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,5120,0.014841778410805596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,5120,0.008984888593355814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,4096,0.00867555538813273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,5120,0.33173155784606934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,4096,0.013783999615245394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,3584,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,4096,0.008002666963471307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,4096,0.32937600877549916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,3584,0.013443555268976422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,3072,0.006324444380071428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,3584,0.007962666451931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,3584,0.3224204381306966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,3072,0.013285333083735572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,2560,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,3072,0.007332444190979004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,3072,0.33402665456136066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,2560,0.012818666795889536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,2048,0.006001777946949005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,2560,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,2560,0.33150577545166016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,2048,0.012440889245933957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,1536,0.005280888742870755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,1536,0.012450666891203986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,2048,0.46696890725029844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,1536,0.3261751068962945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,1024,0.004599110947714912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,1024,0.012070222033394707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,1024,0.005973333285914526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,768,0.003930666794379552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,768,0.012081777883900536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,2048,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,768,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,512,0.003552000141806073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,768,0.3381768862406413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,1536,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,512,0.01181777815024058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,512,0.33428178893195254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,512,0.005751111027267244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,256,0.0032177778581778207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,1024,0.43725956810845273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,256,0.011710222396585675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,128,0.002838222309947014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,256,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,256,0.3324035538567437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,128,0.011456888582971362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,64,0.002817777709828483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1536,128,0.005667555663320754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1536,32,0.002946666752298673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1536,128,0.3447395430670844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,64,0.011729777687125735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,65536,0.024904888537194993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1536,32,0.011063111325105032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,65536,0.05552711089452108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,65536,0.057553774780697294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,65536,0.3788284460703532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,16384,0.022126222650210064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,16384,0.016215110818545025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,12288,0.01239733315176434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,12288,0.018600889378123812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,16384,0.014783999986118741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,12288,0.32939288351270885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,10240,0.01109955542617374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,12288,0.013761777844693927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,10240,0.017874666386180453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,16384,0.3093004491594103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,10240,0.012476444244384766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,10240,0.3218959967295329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,8192,0.010073777702119615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,8192,0.016168889072206285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,8192,0.011019555230935415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,7168,0.00965244405799442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,8192,0.3174684312608507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,7168,0.015811555915408664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,6144,0.008974221845467886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,7168,0.010383110907342698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,7168,0.3344533443450928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,6144,0.01516266663869222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,6144,0.009670221971140968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,5120,0.007337777978844113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,6144,0.28493332862854004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,5120,0.01477688882086012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,5120,0.008979555633332994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,4096,0.007273777491516537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,5120,0.3197902308570014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,4096,0.014118222726715935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,3584,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,4096,0.008299555215570662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,4096,0.33897156185574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,3584,0.013787555197874704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,3072,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,3584,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,3584,0.324694209628635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,3072,0.013366222381591797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,2560,0.005613333235184352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,3072,0.007618666523032718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,3072,0.3265422185262044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,2560,0.01313244468635983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,2048,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,2560,0.0069848886794514125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,2560,0.3311813407474094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,2048,0.013092444174819522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,1536,0.005701333284378052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,2048,0.0069617778062820435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,2048,0.3204960028330485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,1536,0.012668444050682915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,1024,0.004568888909286923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,1536,0.006565333240562015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,1536,0.32096889283921987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,1024,0.012069332930776807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,768,0.004209777961174647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,1024,0.006295111030340195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,1024,0.3312444421980116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,768,0.012079999678664737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,512,0.003536888708670934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,768,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,768,0.32044267654418945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,512,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,256,0.0035093331502543557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,512,0.005593777944644292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,512,0.32890221807691783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,256,0.011699555648697747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,128,0.0031822222388452957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,256,0.005579555614127054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,256,0.32941065894232857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,128,0.011361777782440186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,64,0.0028968888024489083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,1024,128,0.005559111220969095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,1024,32,0.0028773333049482773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,1024,128,0.3028702206081814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,65536,0.021185777253574792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,64,0.011503111157152386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,1024,32,0.010088889135254754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,65536,0.05285066697332594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,16384,0.012994666894276937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,65536,0.057229333453708224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,65536,0.3369777732425266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,16384,0.021310221817758348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,16384,0.32169334093729657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,16384,0.016235555211702984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,12288,0.01092266705301073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,12288,0.018617777360810172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,12288,0.31900355550977916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,10240,0.009463110731707679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,12288,0.01354311075475481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,10240,0.017584888471497428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,10240,0.0123848890264829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,8192,0.008995555341243744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,8192,0.016161777906947665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,10240,0.4610871209038629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,8192,0.3257688946194119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,8192,0.010766222245163388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,7168,0.015841777125994366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,7168,0.3487973213195801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,7168,0.0103502223889033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,6144,0.008071111308203803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,6144,0.015360888507631091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,6144,0.339499553044637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,6144,0.009735110733244155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,5120,0.006636444479227066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,7168,0.008346666892369589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,5120,0.014765333798196582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,4096,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,5120,0.3358711136711969
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,5120,0.009011555876996782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,4096,0.014094221923086377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,3584,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,4096,0.008309333688682979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,4096,0.32583644655015737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,3584,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,3072,0.0069351109365622205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,3584,0.007646222081449296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,3072,0.013169777890046438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,3584,0.39552889929877383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,2560,0.006628444625271692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,3072,0.007640889121426477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,3072,0.3324951065911187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,2560,0.012855110896958245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,2048,0.005277333160241445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,2560,0.007107555866241455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,2560,0.3327626652187771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,2048,0.012497777740160624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,1536,0.004979555391603046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,2048,0.006956444846259222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,2048,0.33538222312927246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,1536,0.012105777859687805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,1024,0.004212444441186057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,1536,0.006296000132958095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,1536,0.32415999306572807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,1024,0.011721777419249216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,768,0.003597333199448056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,1024,0.006280000011126201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,1024,0.33540622393290204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,768,0.011399110986126794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,512,0.003530666646030214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,768,0.005980444451173146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,768,0.33015910784403485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,512,0.011039999624093374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,256,0.0031911111954185697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,512,0.00556355549229516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,512,0.32703200976053876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,256,0.010696888797812991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,128,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,256,0.005642666584915585
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,256,0.3214346567789714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,128,0.010194666683673859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,64,0.0028239999794297745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,768,128,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,768,128,0.30126844512091744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,768,32,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,65536,0.023408000667889912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,64,0.010386666489972008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,768,32,0.010729778144094678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,65536,0.052820444107055664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,16384,0.01070666644308302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,65536,0.05563377671771579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,65536,0.32587289810180664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,16384,0.020988444487253826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,12288,0.00888533310757743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,16384,0.015826667348543804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,16384,0.3358248869578044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,12288,0.01886666648917728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,10240,0.008004444340864817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,12288,0.01346933344999949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,12288,0.3314533233642578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,10240,0.01752088963985443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,8192,0.00835111074977451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,10240,0.012051555845472546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,10240,0.3267902268303765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,8192,0.01643111142847273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,7168,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,8192,0.010984888507260216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,8192,0.32501689592997235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,7168,0.015830222103330824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,6144,0.007288888924651676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,7168,0.010076444182131026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,7168,0.32833067576090497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,6144,0.01531555586391025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,6144,0.009677333136399588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,6144,0.31501245498657227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,5120,0.006958222223652734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,5120,0.01459111107720269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,5120,0.3406204382578532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,5120,0.00904799997806549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,4096,0.007298666569921706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,4096,0.008301333420806462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,3584,0.0069511110583941145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,3584,0.013843555417325763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,3584,0.3377377721998427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,3584,0.007954667011896769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,3072,0.006591111007663939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,4096,0.013798221945762634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,3072,0.013136000268989138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,4096,0.45461331473456484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,2560,0.006333333336644703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,3072,0.34154756863911945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,2560,0.01311644415060679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,2560,0.3412240081363254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,2048,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,2560,0.007095110913117726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,2048,0.012462221913867526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,2048,0.006633777585294511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,1536,0.004590222405062782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,2048,0.38262044058905703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,3072,0.007911110917727152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,1536,0.012497777740160624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,1536,0.006463999963468975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,1024,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,1536,0.3661955462561713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,1024,0.012149333126015134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,768,0.0038648889296584656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,1024,0.006272888845867581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,1024,0.33204976717631024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,768,0.012071111136012606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,512,0.0035208890007601846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,768,0.005976888868543837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,768,0.3583413230048285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,512,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,256,0.003168888803985384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,512,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,512,0.3335626655154758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,256,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,128,0.002903999967707528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,256,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,256,0.3423519929250081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,512,128,0.3025386598375109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,512,128,0.005600000007285013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,64,0.002883555574549569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,512,32,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,64,0.011083555718262991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,65536,0.01477599971824222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,32,0.011024889018800525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,65536,0.050287998384899564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,65536,0.33162400457594127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,16384,0.00905511114332411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,65536,0.04930044545067681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,512,128,0.011372444530328115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,16384,0.02068444424205356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,12288,0.008063111040327284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,16384,0.016519111063745286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,16384,0.3299786779615614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,12288,0.018592000007629395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,10240,0.007679999702506595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,12288,0.013451555536852943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,12288,0.3063226540883382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,10240,0.017261332935757108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,8192,0.00723555518521203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,10240,0.012402666939629449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,10240,0.32663554615444607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,8192,0.016199999385409884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,7168,0.006676444576846228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,8192,0.011078221930397881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,8192,0.32159821192423504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,7168,0.015542222393883599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,6144,0.006303999986913469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,7168,0.010327999790509542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,7168,0.33065954844156903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,6144,0.014832889040311178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,5120,0.006135111053784688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,6144,0.009664889011118147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,6144,0.3177359898885091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,5120,0.014484445254007975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,4096,0.005984000033802456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,5120,0.00903911143541336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,5120,0.3337635464138455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,4096,0.013771555489963956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,3584,0.0062977779242727495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,4096,0.008040000167157915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,4096,0.3167822096082899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,3584,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,3072,0.005859555469618903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,3584,0.007884444461928474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,3584,0.32535023159450954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,3072,0.013208000196350945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,3072,0.00758755538198683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,2560,0.006595555692911148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,2560,0.012795555922720166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,3072,0.3996453285217285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,2048,0.005943999936183293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,2560,0.32227733400132924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,2560,0.0069368887278768755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,2048,0.01218933363755544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,2048,0.006607999818192587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,1536,0.004247111164861255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,2048,0.30860622723897296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,1536,0.012506666282812754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,1024,0.0038951109680864546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,1536,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,1536,0.3558977709876166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,1024,0.012065778175989786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,768,0.003551111039188173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,1024,0.005958222266700532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,1024,0.31365956200493705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,768,0.011773332953453064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,768,0.005899555567238066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,512,0.0032595555401510666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,768,0.34805067380269367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,512,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,256,0.003181333343187968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,512,0.00535999983549118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,512,0.34536088837517637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,256,0.011672889192899069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,128,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,256,0.005600888695981767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,256,0.33917689323425293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,128,0.011377777490350934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,64,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,256,128,0.3087804317474365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,256,128,0.0052382221652401825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,256,32,0.0025964443468385274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,64,0.011407111254003314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,65536,0.014142221874660917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,256,32,0.008993777963850234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,65536,0.04865244362089369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,16384,0.00796888851457172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,65536,0.04798400070932177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,16384,0.021055110626750525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,16384,0.33150577545166016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,16384,0.016238222519556682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,12288,0.006962666908899943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,12288,0.018587556150224473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,12288,0.3285217814975315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,12288,0.013442666994200813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,10240,0.006615110983451207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,10240,0.01758577757411533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,65536,0.3325066566467285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,8192,0.0069306666652361555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,10240,0.3290319972568088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,10240,0.01238222254647149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,8192,0.016148444679048326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,7168,0.006648889018429651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,8192,0.3270204332139757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,7168,0.0155102229780621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,7168,0.010073777702119615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,6144,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,7168,0.31561067369249135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,6144,0.014857777290874057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,6144,0.3232053385840522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,5120,0.005964444329341252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,6144,0.009660444325870937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,8192,0.01071911139620675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,5120,0.014108444253603617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,4096,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,5120,0.008997333546479544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,5120,0.33261066012912327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,4096,0.013464888764752282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,4096,0.008002666963471307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,4096,0.3040844334496392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,3584,0.013520888984203339
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,3584,0.007648000286685095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,3584,0.3018124368455675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,3072,0.006085333310895496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,3072,0.01311911145846049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,3072,0.3461848894755046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,3072,0.007326222128338284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,2560,0.006305777778228124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,2560,0.013064888616402945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,3584,0.006128888991143968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,2048,0.005552000055710475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,2560,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,2048,0.012144000165992312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,2048,0.006622222148709827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,1536,0.004563555535342959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,2048,0.3306151231129964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,1536,0.011730666789743634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,1536,0.31474577056037056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,1536,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,1024,0.003858666867017746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,1024,0.011506666739781698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,2560,0.3349146578047011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,1024,0.322068452835083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,1024,0.005902222047249476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,768,0.011770666473441653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,768,0.32571999231974286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,768,0.006000888844331105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,512,0.0032222221295038858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,512,0.011772444678677453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,512,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,256,0.0028977776981062363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,512,0.4785475730895996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,256,0.011781333221329583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,768,0.0035982223020659555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,128,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,256,0.005574222240183089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,128,0.011034666664070554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,256,0.420364432864719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,64,0.002891555635465516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,8,128,128,0.005617777920431561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,8,128,128,0.30783645311991376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,128,32,0.0029128889242808023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,65536,0.01039644413524204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,64,0.011419555379284753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,16384,0.00701155596309238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,128,32,0.01071999967098236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,12288,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,65536,0.047263112333085805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,10240,0.006257777826653586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,16384,0.020658666888872784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,12288,0.018538667096032035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,8192,0.0069573331210348345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,7168,0.006653333289755716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,10240,0.017232000827789307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,6144,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,8192,0.016176000237464905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,5120,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,7168,0.015481778317027621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,4096,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,6144,0.01482311056719886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,3584,0.005982222242487802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,5120,0.014521777629852295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,3072,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,4096,0.013824000126785703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,3584,0.01351111133893331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,3072,0.013111111190583972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,2048,0.0058959999846087555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,2560,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,1536,0.004565333326657613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,2048,0.012830221818553077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,1024,0.003871110992299186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,1536,0.012445333103338877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,768,0.003568888952334722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,1024,0.012048888537618848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,512,0.003232888877391815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,768,0.011757333245542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,256,0.0029279999434947968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,512,0.011619555453459421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,128,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,256,0.01145866678820716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,2560,0.005952000204059813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,64,0.002524444419476721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,64,32,0.002556444456179937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,128,0.011406222151385413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,65536,0.009373333719041612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,64,0.011039111349317761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,64,32,0.01072799993885888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,65536,0.04829688866933187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,12288,0.006551110910044775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,16384,0.020634666085243225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,10240,0.005913777897755305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,12288,0.018544889158672757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,8192,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,10240,0.017216889394654166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,7168,0.006292444550328785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,8192,0.01649777756796943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,6144,0.0059279998143513995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,7168,0.015558222929636637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,5120,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,6144,0.015142222245534262
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,4096,0.0058444444504049085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,5120,0.014492443866199918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,16384,0.006255110932721033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,3584,0.00589155571328269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,3072,0.005655999812814925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,4096,0.014116444521480136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,2560,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,3584,0.013438222308953604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,2048,0.005609777652555042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,3072,0.013160000244776407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,1536,0.00453333349691497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,2560,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,1024,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,2048,0.012686221963829465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,768,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,1536,0.012483555409643384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,512,0.0032844444115956626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,1024,0.012072888513406118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,256,0.002850666642189026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,768,0.01144266708029641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,128,0.002921777880854077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,512,0.01179377817445331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,64,0.0028195555011431375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,256,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,8,32,32,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,128,0.011067555182509951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,64,0.01202400028705597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,8,32,32,0.01036088913679123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,16384,0.3087048795488146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,16384,0.1894035604265001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,16384,0.14393688572777644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,12288,0.2370471159617106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,12288,0.21048711405860054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,16384,0.3467075559828017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,12288,0.10880888832939996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,10240,0.2008577717675103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,12288,0.339283545811971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,10240,0.16812533802456328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,8192,0.1650302277670966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,10240,0.09247377845976089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,8192,0.13091999954647487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,10240,0.3417920006646051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,7168,0.14428888426886663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,8192,0.0744453337457445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,8192,0.34682310952080625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,7168,0.1182675494088067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,6144,0.127494215965271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,7168,0.06625866889953613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,6144,0.1044702198770311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,7168,0.38951555887858075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,5120,0.10687910848193699
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,6144,0.057336886723836265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,5120,0.08904444509082371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,6144,0.3309395578172472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,4096,0.08404444323645698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,5120,0.05249155561129252
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,5120,0.34090754720899796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,4096,0.039834668238957725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,3584,0.06994933552212186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,4096,0.3470453421274821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,3584,0.06923733154932658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,3584,0.035826666487587824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,3072,0.060812446806165904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,4096,0.07652533054351807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,3584,0.32983554734124076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,3072,0.057108441988627114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,2560,0.051671998368369214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,3072,0.04366844561364916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,2560,0.05378489030732048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,3072,0.32924267980787486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,2048,0.0458586679564582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,2560,0.029061334000693426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,2560,0.32564444012112087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,2048,0.04051288962364197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,1536,0.033013333876927696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,2048,0.030787554052140977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,1536,0.037434667348861694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,2048,0.3385359976026747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,1024,0.02307022280163235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,1536,0.024384000235133704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,1024,0.024482667446136475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,1024,0.018587556150224473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,768,0.018425777554512024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,1024,0.32951643731859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,768,0.023642667465739783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,512,0.013439999686347114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,768,0.31895287831624347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,768,0.015492444237073263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,512,0.019897777173254225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,512,0.013059555656380124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,256,0.008570666942331526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,512,0.32115289900037974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,1536,0.3302968872918023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,128,0.005976888868543837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,256,0.010129777921570672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,128,0.01349511080318027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,256,0.40842223167419434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,64,0.005618666609128316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,65536,128,0.010020444790522257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,65536,128,0.3034702142079671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,65536,32,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,64,0.01456622282663981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,32,0.014256889621416727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,65536,0.31577245394388836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,65536,0.16395555602179632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,16384,0.086352891392178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,65536,256,0.01590577761332194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,16384,0.0787644452518887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,65536,0.1911635531319512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,12288,0.06747555732727051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,16384,0.0461351109875573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,16384,0.3406222131517198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,12288,0.06919822427961561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,10240,0.057709336280822754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,12288,0.03543377916018168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,10240,0.05831911166508993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,12288,0.3542151186201308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,65536,0.3271626631418864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,8192,0.048482666412989296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,10240,0.03357422351837158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,10240,0.33532177077399355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,7168,0.03545866741074456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,8192,0.046915554338031344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,8192,0.027594667341974046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,7168,0.042930665943357676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,8192,0.3344577683342828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,7168,0.024921778175565932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,7168,0.32367200321621364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,6144,0.039057777987586126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,6144,0.32313511106703013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,6144,0.025996444953812495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,5120,0.02683999968899621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,5120,0.03463822272088792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,5120,0.023297778434223596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,4096,0.022648000054889258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,6144,0.03225333491961161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,4096,0.029164443413416546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,4096,0.019179556104871962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,4096,0.33440266715155703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,3584,0.020025778147909377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,3584,0.027644443843099806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,3584,0.34487554762098527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,3584,0.014297778407732645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,3072,0.017584888471497428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,5120,0.33178310924106175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,3072,0.0240657776594162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,3072,0.0177839994430542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,2560,0.015225778023401896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,3072,0.33312177658081055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,2560,0.023079999619060095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,2048,0.01533333294921451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,2560,0.014721777704026965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,2560,0.3441315491994222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,2048,0.019613333874278598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,1536,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,2048,0.012748444245921241
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,2048,0.3322417736053467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,1536,0.018920888503392536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,1024,0.00797155582242542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,1536,0.011154666543006897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,1536,0.32840267817179364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,1024,0.01551555593808492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,768,0.006800000038411882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,1024,0.009090666969617208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,1024,0.33310310045878094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,768,0.01386577801571952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,512,0.004555555681387584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,768,0.008290666672918532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,768,0.333297782474094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,512,0.012884444660610624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,256,0.0032346666687064697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,512,0.007345777418878343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,512,0.32488711675008136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,256,0.01110400011142095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,128,0.003177777760558658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,256,0.006608888920810487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,256,0.3278844356536865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,128,0.011032889286677042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,64,0.0028862222615215513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,16384,128,0.006590222318967183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,16384,128,0.30723733372158474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,16384,32,0.0031759999692440033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,64,0.010355555348926121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,16384,32,0.01072622256146537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,65536,0.22773954603407118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,65536,0.14450400405459934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,16384,0.06435022089216444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,16384,0.06377777788374159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,16384,0.3434568776024713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,65536,0.34831733173794216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,12288,0.051009777519438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,65536,0.14880888991885716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,12288,0.31676443417867023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,12288,0.030960887670516968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,10240,0.04447199900945028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,16384,0.04174044397142199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,10240,0.04896977874967787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,12288,0.06254044506284925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,8192,0.037258668078316584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,10240,0.30103643735249835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,8192,0.041138665543662176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,8192,0.30315110418531627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,8192,0.022587555978033278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,7168,0.03261955579121908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,7168,0.03860266672240363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,7168,0.022244445151752893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,6144,0.03436355458365546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,10240,0.026738666825824316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,6144,0.01825066738658481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,5120,0.02473955518669552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,6144,0.31196800867716473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,5120,0.02944533361328973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,5120,0.31561512417263454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,5120,0.01647644407219357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,4096,0.022115555074479844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,6144,0.028997331857681274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,4096,0.026201777988009985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,3584,0.019144000278578866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,4096,0.31288621160719127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,4096,0.01741155485312144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,3584,0.025012445118692186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,3072,0.016951110627916124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,3584,0.01629333363638984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,3584,0.31091377470228404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,3072,0.022241777843899194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,2560,0.015196444259749519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,3072,0.014099554883109199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,2560,0.020996444755130347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,3072,0.41537067625257706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,7168,0.31614311536153156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,2560,0.32685332828097874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,2048,0.012103111379676394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,2048,0.018606222338146634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,2048,0.32695288128323025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,2048,0.011050666371981302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,1536,0.009072889056470659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,1536,0.01756800048881107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,1536,0.31530843840705025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,1536,0.0097884444726838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,1024,0.007264888948864407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,1024,0.014799111419253878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,2560,0.013206221991115145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,1024,0.3132755491468641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,1024,0.008382221890820397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,768,0.013438222308953604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,768,0.007635555333561367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,768,0.3335715664757623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,512,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,512,0.013078221844302284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,512,0.3376168939802382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,512,0.006959999601046245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,256,0.0032293332947625052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,768,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,256,0.012116444607575735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,128,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,256,0.3124177720811632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,256,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,128,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,64,0.002842666581273079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,12288,128,0.006641777853171031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,12288,128,0.2952337794833713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,64,0.011436444189813403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,12288,32,0.009383111364311641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,65536,0.19098844793107775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,65536,0.13930755191379124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,65536,0.14756088786655003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,12288,32,0.0031511110977994073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,16384,0.05654311180114746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,16384,0.05730666716893514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,65536,0.33884888225131565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,12288,0.04346133271853129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,16384,0.33299377229478627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,16384,0.038176000118255615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,12288,0.050759110185835093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,12288,0.029393778906928167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,10240,0.03802488909827338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,12288,0.3180133236779107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,10240,0.04032977753215366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,8192,0.03122577733463711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,10240,0.02664622167746226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,10240,0.32760622766282826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,8192,0.033626665671666466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,8192,0.022076444493399724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,7168,0.028205331828859117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,8192,0.3262551095750597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,7168,0.03198933270242479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,6144,0.025958221819665697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,7168,0.01995111174053616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,7168,0.34150134192572695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,6144,0.02823822365866767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,6144,0.01754666699303521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,6144,0.32035554779900444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,5120,0.02549422283967336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,5120,0.3238604333665636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,5120,0.01644088824590047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,4096,0.019375110665957134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,4096,0.023023110296991136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,4096,0.31784534454345703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,4096,0.015948444604873657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,3584,0.017106667160987854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,5120,0.021739555729760066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,3584,0.02179199953873952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,3072,0.015948444604873657
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,3584,0.014341332846217685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,3072,0.020261334048377145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,3072,0.31073066923353404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,2560,0.012293332980738746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,2560,0.019279999865425956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,2560,0.3225128915574816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,3584,0.2805253399742974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,2560,0.01144266708029641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,2048,0.011568889021873474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,2048,0.01715555621518029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,1536,0.00870400004916721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,3072,0.013364444176355997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,2048,0.009563555320103964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,2048,0.31736710336473256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,1024,0.006783999916579988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,1536,0.01555377741654714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,1536,0.00869688888390859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,1024,0.014167111780908374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,1536,0.33798755539788145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,768,0.005672889037264719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,1024,0.3066471152835422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,1024,0.00776088900036282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,512,0.004643555730581284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,768,0.013433777623706393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,768,0.007262222468852997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,768,0.3225875430636936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,512,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,256,0.0035137778355015647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,512,0.007176888485749562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,256,0.011744000017642975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,256,0.3220826784769694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,256,0.006602666858169768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,128,0.003192000091075897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,128,0.011443555355072021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,128,0.29639556672838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,10240,128,0.006535111202134027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,10240,32,0.0029173334025674393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,64,0.011064888702498542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,10240,32,0.010055111514197456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,65536,0.15519021617041692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,65536,0.08994844224717881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,65536,0.07822044690450032
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,16384,0.045350223779678345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,16384,0.04936800069279141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,10240,512,0.338939560784234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,16384,0.023028444912698533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,12288,0.0370782216389974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,16384,0.3066364394293891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,12288,0.03858844439188639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,12288,0.31297066476609975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,12288,0.018582221534517076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,10240,0.028359111812379625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,65536,0.3373084333207872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,10240,0.031884445084465876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,8192,0.022633777724372015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,10240,0.3207377857632107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,10240,0.016764443781640794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,8192,0.027465777264700994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,7168,0.020231111182106864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,8192,0.015463999576038785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,8192,0.3135582341088189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,7168,0.02604088849491543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,6144,0.018248889181349013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,7168,0.013516444298956128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,7168,0.32292355431450737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,6144,0.02381688853104909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,5120,0.01586844523747762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,6144,0.013815999858909182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,6144,0.3258764478895399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,5120,0.022279999322361414
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,4096,0.013554666605260638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,5120,0.01146133326821857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,5120,0.32740267117818195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,4096,0.020305777589480083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,3584,0.011903111305501727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,4096,0.010816000401973724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,4096,0.33428356382581925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,3584,0.018802666001849704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,3072,0.010743111371994019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,3584,0.009323555562231276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,3584,0.32476533783806694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,3072,0.018128888474570382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,2560,0.009683555199040307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,3072,0.3288622167375353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,2560,0.017254221770498488
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,2560,0.008632889224423302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,2560,0.329095999399821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,2048,0.009385777844323052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,2048,0.015746666325463187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,2048,0.334990210003323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,2048,0.00795377790927887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,1536,0.007028444773621029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,3072,0.00942844483587477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,1536,0.014377777775128683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,1024,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,1536,0.3333066569434272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,1024,0.01255733354224099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,1024,0.00666933341158761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,1024,0.31786666976081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,768,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,768,0.012784000072214337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,768,0.3340551058451335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,768,0.006294222341643439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,512,0.0035573335157500375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,512,0.012483555409643384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,1536,0.007609777980380588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,512,0.33265688684251576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,256,0.003169777699642711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,512,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,256,0.011386666860845355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,256,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,128,0.002899555489420891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,256,0.3137759897443983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,128,0.011801777614487542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,64,0.002850666642189026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,8192,128,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,8192,32,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,64,0.011196444431940714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,65536,0.1367431084314982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,8192,32,0.010037333601050908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,65536,0.08248977528678046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,16384,0.04115555683771769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,65536,0.07522577709621854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,16384,0.042414221498701304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,16384,0.31098755200703937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,65536,0.3199662102593316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,8192,128,0.2621431085798475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,12288,0.03242133392228021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,16384,0.022749332918061152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,12288,0.0373564428753323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,12288,0.017870222528775532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,12288,0.31313954459296334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,10240,0.03190044562021891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,10240,0.016247999336984422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,10240,0.3144320117102729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,8192,0.02368533280160692
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,8192,0.028207110034094915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,8192,0.3370533254411485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,8192,0.01444533301724328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,7168,0.02107200026512146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,10240,0.028343111276626587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,7168,0.024884444144037034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,6144,0.016544888416926067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,7168,0.012865777644846173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,7168,0.33103378613789874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,6144,0.022729777627521094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,6144,0.012422222230169507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,5120,0.014477334088749357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,6144,0.3302862114376492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,5120,0.021433777279324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,4096,0.013349333571063148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,5120,0.01109955542617374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,5120,0.32893066936069065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,4096,0.019855999284320407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,4096,0.008984888593355814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,4096,0.3507644335428874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,3584,0.01958577831586202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,3584,0.008714666797055138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,3584,0.3071991072760688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,3072,0.009953777823183272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,3072,0.017342221405771043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,3072,0.3285164568159315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,3072,0.008610666626029545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,2560,0.008626666333940294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,3584,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,2560,0.016566221912701923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,2048,0.008753777378135258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,2560,0.3165386782752143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,2560,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,2048,0.014840889308187695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,1536,0.0069120000633928515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,2048,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,2048,0.3093075487348768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,1536,0.013801777528391944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,1024,0.0047102222839991255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,1536,0.006599111275540457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,1536,0.3254275586869982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,1024,0.012437333663304647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,768,0.0041422223051389056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,1024,0.31818845536973744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,768,0.012458666331238218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,768,0.006057777752478917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,512,0.003583111282851961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,512,0.012050666742854647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,512,0.33809600936041934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,1024,0.006594666590293248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,256,0.0031822222388452957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,512,0.005959110955397288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,256,0.011764444410800934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,768,0.3112186590830485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,256,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,256,0.33128711912367076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,128,0.002854222224818336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,128,0.011800888511869641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,7168,128,0.005568888866239124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,64,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,7168,32,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,7168,128,0.2987511157989502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,64,0.009691555466916826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,7168,32,0.01012711144155926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,65536,0.07536355654398601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,65536,0.07136533657709758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,16384,0.03644266724586487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,65536,0.3082462151845296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,16384,0.039665778477986656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,65536,0.11933688322703044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,16384,0.021669333179791767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,16384,0.3113226625654432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,12288,0.03001333276430766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,12288,0.036991112762027316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,10240,0.024481778343518574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,12288,0.01738133364253574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,12288,0.3110915554894341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,10240,0.030400001340442236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,8192,0.02083288960986667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,8192,0.027057778504159715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,10240,0.4703182114495172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,8192,0.013934221532609729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,8192,0.32155466079711914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,7168,0.018883556127548218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,7168,0.024362666739357844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,7168,0.01275466630856196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,6144,0.017293334007263184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,7168,0.30436621771918404
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,6144,0.022597332795461018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,6144,0.34104177686903214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,5120,0.015153777268197803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,6144,0.011576000187132092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,10240,0.015658666690190632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,5120,0.02106488909986284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,5120,0.01035733355416192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,4096,0.012038222617573209
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,5120,0.37519465552435977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,4096,0.018950222267044913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,3584,0.010246222217877706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,4096,0.30555465486314565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,4096,0.00906311141120063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,3584,0.018518222702874076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,3072,0.009342222577995723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,3584,0.29872534010145396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,3072,0.017158221867349412
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,3072,0.008018666671382057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,3072,0.30737243758307564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,2560,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,2560,0.015855110353893705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,2560,0.34778489006890195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,2560,0.007276444799370236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,2048,0.007699555820888943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,3584,0.008672888908121321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,2048,0.014821334017647637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,1536,0.0063386667105886675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,2048,0.335115565194024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,2048,0.007065777977307637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,1536,0.013789333403110504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,1024,0.0042364444169733255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,1536,0.006688000013430913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,1536,0.32747112380133736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,1024,0.012173333101802401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,768,0.003859555555714501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,1024,0.006326222171386083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,1024,0.34643999735514325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,768,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,512,0.00350222239891688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,768,0.006137777947717243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,768,0.3316728803846571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,512,0.011384000380833944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,256,0.003236444460021125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,512,0.005939555664857228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,512,0.3397288852267795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,256,0.01070844464831882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,128,0.0028595555987623003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,256,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,256,0.34152266714308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,6144,128,0.005921777751710679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,6144,128,0.3083439932929145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,64,0.011051555474599203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,32,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,32,0.01035644445154402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,65536,0.1023813353644477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,6144,128,0.010695111420419482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,65536,0.07012444734573364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,6144,64,0.0029128889242808023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,16384,0.032319999403423734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,65536,0.06892800331115723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,16384,0.04066133168008592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,65536,0.33869067827860516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,12288,0.025378665990299646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,16384,0.34637777010599774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,12288,0.03429866830507914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,12288,0.3384542200300429
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,12288,0.016914667354689706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,10240,0.02202044427394867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,10240,0.028008000718222723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,10240,0.3338293234507243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,10240,0.015223999818166098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,16384,0.021039111746682063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,8192,0.02510844502184126
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,8192,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,7168,0.01700088878472646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,8192,0.3338035477532281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,7168,0.02331466641690996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,7168,0.012074666718641916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,7168,0.33613332112630206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,6144,0.015537778536478678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,8192,0.019241778386963736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,6144,0.021731555461883545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,5120,0.013889777991506787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,6144,0.01110755569405026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,6144,0.32536265585157603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,5120,0.020566221740510728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,4096,0.011065777805116443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,5120,0.009787555370065901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,5120,0.33384089999728733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,4096,0.01854933301607768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,3584,0.009677333136399588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,4096,0.32694045702616376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,4096,0.008364444805516137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,3584,0.017272000511487324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,3072,0.008752000000741746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,3584,0.007999999655617608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,3584,0.3120373355017768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,3072,0.015625778171751235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,2560,0.007976888782448238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,3072,0.007638221813572778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,3072,0.3352639940049913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,2560,0.014864888456132678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,2048,0.008029333419269985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,2560,0.28969422976175946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,2560,0.007327111230956183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,2048,0.00693866651919153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,2048,0.013847110999955071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,1536,0.00526311124364535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,2048,0.3045724497901069
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,1536,0.01351199961370892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,1536,0.006607111129495833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,1536,0.3351084391276042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,1024,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,1024,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,1024,0.32849688000149196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,768,0.0038720000949170855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,768,0.012459555433856117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,768,0.32861600981818306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,768,0.006274666637182236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,1024,0.003947555604908201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,512,0.012133333418104382
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,512,0.005968889014588461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,256,0.003273777870668305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,512,0.3098364406161838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,256,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,256,0.30409688419765896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,256,0.0058408888677755995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,128,0.003173333282272021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,512,0.003522666792074839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,128,0.01181688904762268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,64,0.002891555635465516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,5120,128,0.2915591133965386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,5120,32,0.0028008888992998335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,64,0.011128000087208219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,5120,32,0.009708444277445475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,65536,0.06555110878414579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,5120,128,0.005622222191757626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,65536,0.32778310775756836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,16384,0.026685333914226953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,65536,0.08217066526412964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,65536,0.06659644179873996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,16384,0.033302221033308245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,12288,0.022050667140218947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,16384,0.02016888890001509
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,16384,0.33933954768710667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,12288,0.030760000149408977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,10240,0.018554665976100497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,12288,0.33204176690843373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,12288,0.01632355484697554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,10240,0.02637422250376807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,10240,0.014848889576064216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,10240,0.33625422583685977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,8192,0.022646221849653456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,8192,0.012835555606418185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,7168,0.014500444134076437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,8192,0.30139732360839844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,7168,0.02128177715672387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,7168,0.3146515687306722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,6144,0.013010666602187686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,7168,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,8192,0.016864000095261466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,6144,0.02029511166943444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,5120,0.012317332956526013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,6144,0.010390222072601318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,6144,0.34142398834228516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,5120,0.019387554791238572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,4096,0.011996444728639392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,5120,0.00904622260067198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,5120,0.3187955485449897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,4096,0.01722666621208191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,3584,0.011141333315107556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,4096,0.008048888709810045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,4096,0.3267751004960802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,3584,0.015221332510312399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,3072,0.010043555663691627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,3584,0.007699555820888943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,3584,0.32080356280008954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,3072,0.01481244464715322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,2560,0.008868444297048781
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,3072,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,3072,0.3288817670610216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,2560,0.01385155568520228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,2048,0.008271999657154083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,2560,0.007280000381999546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,2560,0.32339821921454537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,2048,0.013394667042626275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,1536,0.006046222315894232
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,2048,0.006937777830494775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,2048,0.3259929021199544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,1536,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,1024,0.005208888815508948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,1536,0.006682666639486949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,1536,0.3255457878112793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,1024,0.012063110868136087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,768,0.004589333302444882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,1024,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,1024,0.32454310523139107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,512,0.003957333250178231
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,768,0.012056888805495368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,768,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,512,0.011783999701340994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,512,0.32743822203742134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,768,0.4781493080986871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,512,0.005963555640644497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,256,0.0032426667296224167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,256,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,256,0.010425777898894416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,128,0.0029004443850782183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,256,0.33945332633124453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,128,0.010939555863539377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,4096,128,0.005790222022268508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,64,0.0029120000286234748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,4096,128,0.314538664287991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,4096,32,0.0028888889484935333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,65536,0.07435466845830281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,64,0.011042666931947073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,4096,32,0.009505777723259395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,65536,0.06308622492684259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,16384,0.023406222462654114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,65536,0.33268711302015513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,16384,0.03127022253142463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,16384,0.019889778561062284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,16384,0.32208177778455943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,12288,0.01908533275127411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,12288,0.02788355615403917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,12288,0.3421333365970188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,12288,0.016215110818545025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,10240,0.01648977730009291
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,65536,0.06560533576541476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,10240,0.024401777320437964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,10240,0.014656000667148165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,8192,0.014934221903483072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,10240,0.3291155497233073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,8192,0.022043555974960327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,7168,0.01346933344999949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,8192,0.01216533366176817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,8192,0.33660888671875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,7168,0.021026665965716045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,6144,0.012278222375445895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,7168,0.011367110742463006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,7168,0.3265964455074734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,6144,0.01926488843229082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,5120,0.011431111229790581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,6144,0.010686222049925061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,6144,0.32745244767930776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,5120,0.01814311080508762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,5120,0.009577777650621202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,4096,0.011530666715568967
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,5120,0.32187289661831325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,4096,0.015999999311235216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,3584,0.01143200033240848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,4096,0.007968000239796108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,4096,0.3126311037275526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,3584,0.015103111664454142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,3072,0.010067555639478896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,3584,0.2924871179792616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,3072,0.014820444915029736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,3072,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,3072,0.29768265618218315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,2560,0.008420444197124904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,2560,0.013810666898886362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,2560,0.33982311354743105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,2048,0.007321777443091075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,3584,0.008184889124499427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,2048,0.013438222308953604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,2048,0.006744888921578725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,1536,0.006324444380071428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,2048,0.3294631110297309
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,1536,0.01316177762216992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,1536,0.320578654607137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,1536,0.006667555620272954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,1024,0.004922666483455234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,2560,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,1024,0.012440000143316058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,768,0.004250666747490565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,1024,0.0063031112982167145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,768,0.012072000238630505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,768,0.006001777946949005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,768,0.31940089331732857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,512,0.011095110740926532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,512,0.3179813226064046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,1024,0.32852265569898814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,512,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,256,0.0031715554909573663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,256,0.010693333215183683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,512,0.003938666648334927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,256,0.3164915508694119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,128,0.0029102222373088202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,256,0.005604444278611078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,128,0.011427555647161273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3584,128,0.005560889012283749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,64,0.002899555489420891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3584,32,0.0029066666546795103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3584,128,0.30459022521972656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,64,0.009644444617960189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3584,32,0.009376000199053023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,65536,0.06100800302293566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,16384,0.020480889413091872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,65536,0.06457866562737359
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,65536,0.3331262270609538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,16384,0.029264890485339697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,16384,0.31762221124437123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,12288,0.017889777819315594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,65536,0.06514844629499647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,12288,0.027790221903059218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,12288,0.015961777832772996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,12288,0.34081512027316624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,10240,0.014984000060293408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,10240,0.02268444498380025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,10240,0.3225946691301134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,10240,0.014164444473054675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,8192,0.013052444491121503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,8192,0.020655999581019085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,16384,0.01977511081430647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,7168,0.012223111258612739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,8192,0.3159342341952854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,8192,0.011765333513418833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,7168,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,6144,0.012159999873903064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,7168,0.3130551179250081
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,6144,0.01862844493654039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,6144,0.3127937846713596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,6144,0.009700444837411245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,5120,0.010372444159454769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,5120,0.016232000456915963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,5120,0.32229333453708225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,5120,0.009033777647548253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,7168,0.020374221934212577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,4096,0.014873777826627096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,4096,0.008031999899281396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,3584,0.009853333234786987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,4096,0.3970044453938802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,3584,0.014538667268223233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,3584,0.28994401295979816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,3584,0.007747555772463481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,3072,0.008663111262851292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,4096,0.010440000229411654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,3072,0.014165333575672574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,2560,0.0075288886825243635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,3072,0.007426666716734569
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,3072,0.30512089199490017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,2560,0.01352533366945055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,2048,0.006340444501903322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,2560,0.007298666569921706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,2560,0.3144222100575765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,2048,0.01313244468635983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,1536,0.006245333287451003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,2048,0.007024889190991719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,1536,0.012847111456924014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,1536,0.0069306666652361555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,1536,0.3202008936140272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,1024,0.004852444347408083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,1024,0.011439999772442712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,1024,0.32868888643052846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,1024,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,768,0.004551110996140374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,768,0.011395555403497485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,2048,0.32216978073120117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,768,0.3212675518459744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,512,0.003867555409669876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,768,0.005920889063013925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,512,0.005613333235184352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,256,0.003567999849716822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,512,0.30682489607069224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,256,0.010718222293588849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,256,0.005600888695981767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,256,0.3112604353162977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,128,0.0032239999208185407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,128,0.01143555591503779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,3072,128,0.29957421620686847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,3072,128,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,64,0.00253866674999396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,3072,32,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,512,0.011807999677128263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,64,0.0103502223889033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,3072,32,0.008815111385451423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,65536,0.0596311092376709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,65536,0.06123377879460653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,65536,0.31979555553860134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,16384,0.01959022217326694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,16384,0.0274000002278222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,16384,0.33507998784383136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,12288,0.015824000040690105
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,65536,0.05754488706588745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,12288,0.02552622225549486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,12288,0.015140444040298462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,12288,0.3031795554690891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,10240,0.013653332988421122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,10240,0.022107554806603327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,8192,0.01201244443655014
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,10240,0.3294515609741211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,10240,0.013601777454217276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,8192,0.01994311147265964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,7168,0.011367110742463006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,16384,0.01938222183121575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,8192,0.32371022966172963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,8192,0.011225777367750803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,7168,0.018890667292806838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,6144,0.01145511120557785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,7168,0.010753778119881948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,7168,0.3252497778998481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,6144,0.017260443833139207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,5120,0.009780444204807281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,6144,0.009767110976907942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,6144,0.4025591214497884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,5120,0.009104889300134447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,5120,0.3118071026272244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,4096,0.010244444840484196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,4096,0.01480888823668162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,4096,0.007948444121413762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,4096,0.3737653361426459
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,3584,0.008301333420806462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,3584,0.014155555102560254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,3584,0.3217466672261556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,3072,0.007465778125656976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,3584,0.007962666451931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,3072,0.013840888937314352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,3072,0.007629333270920648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,2560,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,2560,0.0131217779384719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,5120,0.015867556134859722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,2560,0.3231004344092475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,2048,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,2560,0.00703022215101454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,2048,0.01274311128589842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,3072,0.35949688487582737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,2048,0.0069671107663048645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,1536,0.005677333308590784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,2048,0.3394924534691705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,1536,0.012077333198653327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,1536,0.006676444576846228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,1024,0.004612444589535396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,1536,0.340664890077379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,1024,0.012113778127564324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,768,0.004509333521127701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,1024,0.006224000205596288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,1024,0.32301333215501576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,768,0.011763555308183035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,512,0.003899555653333664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,768,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,768,0.33574843406677246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,512,0.011710222396585675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,256,0.0035297779573334586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,512,0.3249991205003527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,256,0.011053333679835001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,128,0.0033119999700122406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,256,0.005588444570700328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,256,0.3288568920559353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,128,0.011761777930789523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,64,0.002790222151411904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2560,128,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2560,128,0.29732267061869305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2560,32,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,64,0.011006222003036074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,65536,0.04807555675506592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2560,32,0.008674666285514832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,65536,0.057484441333346896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,16384,0.016199111938476562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,65536,0.05875644418928358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,65536,0.3425697750515408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,16384,0.025007999605602686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,12288,0.013060444758998023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,16384,0.3304453425937229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,16384,0.018953777021831937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,12288,0.022824888428052265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,10240,0.011798222031858234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,12288,0.014144000079896716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,12288,0.32380355728997123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,10240,0.020322667227851022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,8192,0.01073688848151101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,10240,0.012633778154850006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,10240,0.34458133909437394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,8192,0.018168000711335074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,7168,0.010399999717871347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,8192,0.010862222148312463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,8192,0.31436533398098415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,7168,0.016521778371598985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,6144,0.010694222317801582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,7168,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,7168,0.3394577768113878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,6144,0.015407111909654407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,5120,0.00907555553648207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,6144,0.009711999860074785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,6144,0.29911555184258354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,5120,0.014594667487674289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,4096,0.009719111025333405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,5120,0.009072889056470659
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,5120,0.30365334616767037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,4096,0.01384800010257297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,3584,0.008972444468074376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,4096,0.00832799987660514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,4096,0.30487288369072807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,3584,0.013457777599493662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,3072,0.006575999988449945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,3584,0.007931555310885111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,3584,0.3174435562557644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,3072,0.013462222284740873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,2560,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,3072,0.007659555309348636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,3072,0.3153617646959093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,2560,0.013086222112178802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,2560,0.007231111327807109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,2560,0.3106951183742947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,2048,0.006751110984219446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,2048,0.01276177747382058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,2048,0.34152534272935653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,2048,0.006956444846259222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,1536,0.006310222049554189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,1536,0.012076444096035428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,1024,0.0053377776510185665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,1536,0.3287173377143012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,1536,0.006615110983451207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,1024,0.005984888722499211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,1024,0.012216889195972018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,768,0.004184000194072723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,1024,0.32792533768547905
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,768,0.01185155577129788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,768,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,512,0.003956444561481476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,768,0.3308444552951389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,512,0.011688888900809817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,512,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,512,0.4130382272932265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,256,0.010372444159454769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,256,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,128,0.003208000005947219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,256,0.3813653257158067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,128,0.010575110713640848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,2048,128,0.005588444570700328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,2048,128,0.33205689324273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,256,0.0035751110149754416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,32,0.0029128889242808023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,64,0.01039644413524204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,65536,0.03463733196258545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,2048,32,0.010182222558392419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,65536,0.05563466416464912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,65536,0.3389866616990831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,65536,0.058224002520243325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,16384,0.013493333425786761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,16384,0.024199111594094172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,16384,0.33229509989420575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,16384,0.018221333622932434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,12288,0.01110311100880305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,12288,0.021712000171343487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,12288,0.3273715443081326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,12288,0.013778666655222574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,10240,0.013759111364682516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,10240,0.018971555762820773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,10240,0.31387021806504994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,8192,0.012703110774358114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,10240,0.01239377756913503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,2048,64,0.0028515555378463534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,8192,0.01682844426896837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,8192,0.01073155552148819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,7168,0.011371555427710215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,8192,0.3390195634629991
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,7168,0.01553066737122006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,6144,0.010565333068370819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,7168,0.010427555276287926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,7168,0.32155201170179576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,6144,0.015134221977657743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,5120,0.009547555612193214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,6144,0.009700444837411245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,6144,0.3343057897355821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,5120,0.014560000763999091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,4096,0.008630221916569604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,5120,0.008983111215962304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,5120,0.32415734397040474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,4096,0.013743110828929476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,4096,0.007958222594526079
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,4096,0.3309217823876275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,3584,0.013832888669437833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,3584,0.007663999994595845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,3584,0.32167379061381024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,3072,0.006362666686375936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,3072,0.013442666994200813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,3072,0.33872267935011124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,2560,0.0059573331640826324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,3584,0.006670222100284364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,2560,0.01309866706530253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,2560,0.007253333098358578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,2048,0.00591466658645206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,2560,0.3322835498385959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,2048,0.01275288893116845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,2048,0.006726222319735422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,2048,0.3172649012671577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,1536,0.005224888937340842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,3072,0.007638221813572778
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,1536,0.012049777640236748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,1024,0.004197333421972063
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,1536,0.006678222368160884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,1536,0.33458399772644043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,1024,0.011141333315107556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,768,0.004234666625658671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,1024,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,768,0.0110871113008923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,512,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,768,0.005987555616431766
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,512,0.01074399964676963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,768,0.45649512608846027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,512,0.3236151006486681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,256,0.0032302221904198327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,512,0.005670222143332164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,256,0.011128889189826118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,256,0.2880275514390734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,256,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,128,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,1024,0.306130674150255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,128,0.011695110963450538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,64,0.002806222273243798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1536,128,0.005584888988071018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1536,128,0.29421689775254994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1536,32,0.002895111011134254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,65536,0.02497244377930959
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,64,0.011037333144081963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1536,32,0.00943377779589759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,65536,0.0535715553495619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,16384,0.01463644372092353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,65536,0.05753866831461588
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,65536,0.33367466926574707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,16384,0.022688888841205176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,12288,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,16384,0.016213332613309223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,16384,0.3262462351057264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,12288,0.018927110566033255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,10240,0.011130666567219628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,12288,0.32898489634195965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,10240,0.01752622259987725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,10240,0.01240355521440506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,10240,0.33547200096978086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,8192,0.009992889232105678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,8192,0.016159999701711867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,8192,0.3237333297729492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,8192,0.01073688848151101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,7168,0.009003555609120263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,12288,0.013850666582584381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,7168,0.015558222929636637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,6144,0.008458666503429413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,7168,0.3182835578918457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,7168,0.010216889282067617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,6144,0.015244444211324057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,5120,0.007315555380450354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,6144,0.009719111025333405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,6144,0.32073243459065753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,5120,0.014486221803559197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,4096,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,5120,0.009018667042255402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,5120,0.33035021358066136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,4096,0.013742222554153867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,3584,0.007253333098358578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,4096,0.008295999632941352
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,4096,0.324290672938029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,3584,0.013521778086821238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,3072,0.006303999986913469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,3584,0.007963555554548899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,3584,0.3270222345987956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,3072,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,2560,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,3072,0.007598222129874759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,3072,0.32485511567857533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,2560,0.013037333057986366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,2048,0.006633777585294511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,2560,0.007258666886223688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,2560,0.3283280001746284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,2048,0.012715555727481842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,1536,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,2048,0.006978666616810693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,2048,0.32467910978529185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,1536,0.01241422196229299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,1024,0.005290666802061929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,1536,0.006589333216349284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,1536,0.3266346719529894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,1024,0.012118221984969245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,768,0.004629333400064045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,1024,0.006226666685607698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,1024,0.32781778441535103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,768,0.011766222616036734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,512,0.0038844446341196695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,768,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,768,0.32886044184366864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,512,0.01148977792925305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,256,0.003536000019974179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,512,0.00564444437623024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,512,0.3210444450378418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,256,0.011043555206722684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,128,0.003175111073586676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,256,0.005568000177542369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,256,0.3261360062493218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,128,0.011821333732869891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,64,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,1024,128,0.005567111074924469
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,1024,128,0.297980440987481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,1024,32,0.002895999906791581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,65536,0.02018311123053233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,64,0.011009777585665384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,1024,32,0.009443555441167619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,65536,0.052390221092436046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,16384,0.01278222186697854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,65536,0.05791999896367391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,65536,0.3394942283630371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,16384,0.020922667450375024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,16384,0.3329297701517741
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,16384,0.016494222813182406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,12288,0.010428444378905825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,12288,0.018446221947669983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,12288,0.3341244326697455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,12288,0.013803555733627744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,10240,0.009310222334331935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,10240,0.017229333519935608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,10240,0.3343937661912706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,10240,0.012436444560686747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,8192,0.009127111070685917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,8192,0.016125332978036668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,7168,0.008349333372380998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,8192,0.3260115517510308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,8192,0.010447111394670276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,7168,0.015462223026487561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,6144,0.00793422261873881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,7168,0.010101333260536194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,7168,0.32792356279161244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,6144,0.014831110835075378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,5120,0.006647999915811751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,6144,0.009635555247465769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,5120,0.014474666780895658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,6144,0.4425324334038629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,4096,0.007648888561460707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,5120,0.3356000052558051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,5120,0.00905777762333552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,4096,0.008065777520338694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,4096,0.013968888256284924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,3584,0.007283555964628856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,4096,0.3299608760409885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,3584,0.013501333693663279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,3072,0.006612444503439798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,3584,0.0076657773719893555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,3584,0.33962400754292804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,3072,0.01313688854376475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,3072,0.007589333587222629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,3072,0.32899289660983616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,2560,0.012413333687517377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,2560,0.33404265509711367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,2560,0.007368889119890001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,2048,0.005221333354711533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,2048,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,2048,0.3220000002119276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,2048,0.0069919998447100324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,1536,0.005667555663320754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,1536,0.012126222252845764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,1536,0.32994045151604545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,1024,0.0042444442709287005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,1024,0.012060444388124677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,1024,0.32167911529541016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,1024,0.006253333141406377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,768,0.003960000144110786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,2560,0.006619555668698416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,768,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,1536,0.006576889091067844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,512,0.0035173334181308746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,768,0.30742398897806805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,768,0.005960888746711943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,512,0.011352889239788055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,256,0.003149333306484752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,512,0.005888888819350137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,512,0.29224356015523273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,256,0.011078221930397881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,128,0.00315022220214208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,256,0.005281777845488654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,256,0.3291200002034505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,128,0.011174221833546957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,64,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,768,128,0.005630222045713001
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,768,32,0.0028684443483750024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,768,128,0.28293776512145996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,65536,0.022658665974934895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,64,0.010066666536860997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,768,32,0.010079999764760336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,65536,0.05117422342300415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,16384,0.010472888747851053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,65536,0.055592887931399874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,65536,0.34744445482889813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,16384,0.020297777321603563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,12288,0.008664000365469191
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,16384,0.016354666815863717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,16384,0.3330889013078478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,12288,0.018248889181349013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,10240,0.007644444704055786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,12288,0.013459555804729462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,12288,0.3422702153523763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,10240,0.016954667038387723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,8192,0.007992888490358988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,10240,0.012107555237081317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,10240,0.3330133226182726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,8192,0.015826667348543804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,7168,0.007658667034573025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,8192,0.010782221953074137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,8192,0.33132889535692006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,7168,0.015180443723996481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,6144,0.0069395556218094295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,7168,0.010010666317409938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,7168,0.3444844351874457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,6144,0.01479822231663598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,6144,0.009376000199053023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,6144,0.3392213450537787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,5120,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,5120,0.014507555299335055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,4096,0.007328888608349695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,5120,0.00905866672595342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,5120,0.329785770840115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,4096,0.013521778086821238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,3584,0.0069360000391801195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,4096,0.007974222302436829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,3584,0.01348888874053955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,4096,0.4532151222229004
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,3072,0.006974221931563483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,3584,0.007696888513035244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,3072,0.013449778159459433
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,3072,0.3352399932013617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,3072,0.007460444337791867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,2560,0.0064035554726918536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,2560,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,2560,0.33107733726501465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,2048,0.005950222412745158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,2560,0.007263110743628607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,3584,0.3314346737331814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,2048,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,1536,0.004590222405062782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,2048,0.006635555376609166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,2048,0.3642444345686171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,1536,0.012443555725945367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,1024,0.003903999924659729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,1536,0.006610666712125142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,1536,0.3362773259480794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,1024,0.012128888732857175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,768,0.003677333394686381
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,1024,0.005746666755941179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,1024,0.3403724564446344
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,768,0.011791999969217511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,512,0.0035253332720862497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,768,0.005934222290913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,768,0.32293423016866046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,512,0.011415111521879831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,256,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,512,0.005241777747869492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,512,0.33591198921203613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,256,0.010988444089889526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,128,0.003149333306484752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,256,0.005590222362014983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,256,0.31300444073147243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,128,0.011694221860832639
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,64,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,512,128,0.005251555393139522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,512,32,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,512,128,0.30491288503011066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,64,0.01110222190618515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,65536,0.014848889576064216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,512,32,0.00835111074977451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,65536,0.049991998407575816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,16384,0.008711111214425828
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,65536,0.04861333303981357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,65536,0.3275928762223986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,16384,0.020638222495714825
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,12288,0.007995555798212687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,16384,0.016546666622161865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,16384,0.3343671162923177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,12288,0.018184888694021437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,10240,0.007607111500369177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,12288,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,10240,0.017265778448846605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,10240,0.3281911214192708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,8192,0.006981333096822103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,8192,0.015839111473825242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,8192,0.280823998981052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,12288,0.3232453399234348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,8192,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,7168,0.014935111006100973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,10240,0.012426666915416718
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,7168,0.010422222316265106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,7168,0.3137679894765218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,6144,0.01479111115137736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,7168,0.00664533343580034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,6144,0.3088124328189426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,5120,0.005991111199061076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,6144,0.009678222239017487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,5120,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,5120,0.008654221892356873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,4096,0.005966222120655908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,4096,0.01314222233162986
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,4096,0.32595022519429523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,3584,0.0064382221963670515
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,4096,0.007907555335097844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,3584,0.013819555441538492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,6144,0.006298666612969504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,5120,0.32121867603725857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,3584,0.32151733504401314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,3072,0.006238222122192383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,3584,0.007656888829337225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,2560,0.006243555496136348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,3072,0.013378666506873237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,3072,0.007312888900438945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,2560,0.012429333395428128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,3072,0.31629599465264213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,2048,0.00600622221827507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,2560,0.007277333074145847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,2560,0.3114062150319417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,2048,0.012127999630239276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,1536,0.004569777597983678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,2048,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,2048,0.3315395514170329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,1536,0.011741333537631564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,1024,0.0038684445122877755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,1536,0.32370132870144314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,1024,0.011044444309340583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,1024,0.30752621756659615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,768,0.003527111063400904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,768,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,768,0.3152782122294108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,1536,0.006262222097979651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,512,0.003444444388151169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,768,0.005610666755172942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,512,0.01142133358452055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,1024,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,256,0.0028515555378463534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,512,0.326836453543769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,512,0.00556355549229516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,256,0.010372444159454769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,128,0.0028400001012616684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,256,0.00553955551650789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,256,0.31852444012959796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,128,0.010071111222108206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,64,0.0026782221264309357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,256,128,0.005587555468082428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,256,32,0.002579555536309878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,256,128,0.31308889389038086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,65536,0.014152000347773233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,64,0.009009777671760982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,256,32,0.009558222360081142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,65536,0.04718044400215149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,16384,0.00795022232664956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,65536,0.04899200134807163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,65536,0.33839021788703066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,16384,0.020423110988405015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,12288,0.007626666790909237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,16384,0.01588800052801768
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,16384,0.3346746762593587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,12288,0.01826400061448415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,10240,0.007275555696752336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,12288,0.013796444568369122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,12288,0.32250934176974827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,10240,0.01707466608948178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,8192,0.006630222416586346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,10240,0.011714666253990598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,10240,0.32349154684278697
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,8192,0.01590133375591702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,7168,0.006659555352396435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,8192,0.010768888725174798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,8192,0.3132755491468641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,7168,0.01518311103185018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,6144,0.006714666469229593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,7168,0.010030222435792288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,7168,0.3345057699415419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,6144,0.014511111709806653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,6144,0.009749333063761393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,6144,0.3192515638139513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,5120,0.014167999227841696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,5120,0.30934132470024955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,5120,0.008967111508051554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,4096,0.00563822231358952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,4096,0.013797333670987023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,4096,0.31930843989054364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,4096,0.008018666671382057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,3584,0.006260444306664997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,5120,0.006263999889294307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,3584,0.3101777765485975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,3072,0.0059075554211934405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,3072,0.012406222522258759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,3072,0.3135760095384386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,3072,0.007275555696752336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,2560,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,3584,0.013783111340469785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,2560,0.012028444144460889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,3584,0.007715555528799693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,2048,0.005608888963858287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,2560,0.28686756557888454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,2560,0.0069644442862934535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,2048,0.012416889270146688
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,1536,0.00453333349691497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,2048,0.30928089883592397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,1536,0.012098666694429187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,1536,0.006581333362393909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,1024,0.0038524443904558816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,1024,0.011767110890812345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,1024,0.332261323928833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,2048,0.00666844430896971
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,768,0.0035048888789282907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,1024,0.006314666734801398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,768,0.010747555229398938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,512,0.0032604444358083936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,768,0.005991111199061076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,768,0.30433866712782115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,1536,0.30315110418531627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,512,0.010688888529936472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,256,0.002895111011134254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,512,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,512,0.4289582305484348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,256,0.010016000105275048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,128,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,256,0.005605333381228977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,256,0.3200204372406006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,128,0.010029333333174387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,64,0.002504000026318762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,4,128,128,0.005640888793600931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,128,32,0.0025191110455327565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,64,0.009088888764381409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8_block,4,128,128,0.3384693463643392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,65536,0.010022222167915767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,16384,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,128,32,0.009347555538018545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,12288,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,65536,0.046557333734300405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,16384,0.020260444945759244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,10240,0.006228444476922353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,8192,0.006648889018429651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,12288,0.018522666560279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,7168,0.006302222195598815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,10240,0.016857778032620747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,6144,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,8192,0.015836444165971544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,7168,0.015477332803938123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,5120,0.006082666830884085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,4096,0.005623111294375525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,6144,0.014544000228246054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,3584,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,5120,0.014553777045673795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,3072,0.006112888869312074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,4096,0.013537777794731988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,2560,0.005973333285914526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,3584,0.013105777402718862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,2048,0.0052577778697013855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,3072,0.013506666653686099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,1536,0.004538666870858935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,2560,0.012869333227475485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,1024,0.00387999994887246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,2048,0.012100444071822695
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,768,0.003536000019974179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,1536,0.012413333687517377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,512,0.0035582222044467926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,1024,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,768,0.01073155552148819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,512,0.011510222322411008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,128,0.0028560000161329904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,256,0.010055999788973067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,64,0.0028275555620590844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,128,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,32,0.00282577777074443
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,64,0.00869511150651508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,65536,0.009024000002278222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,64,32,0.009655111365848118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,16384,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,65536,0.04632711079385546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,12288,0.0058986664646201665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,16384,0.020628444022602506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,10240,0.00600000015563435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,64,256,0.0029155556112527847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,12288,0.01790488925245073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,8192,0.005597333527273602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,7168,0.005903999838564131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,10240,0.017260443833139207
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,6144,0.006633777585294511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,8192,0.015464888678656684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,5120,0.005569777968857024
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,7168,0.015488000379668342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,4096,0.0052586665583981415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,6144,0.014854222536087036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,3584,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,5120,0.014134221606784396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,3072,0.0052782222628593445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,4096,0.01350488927629259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,2560,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,3584,0.013459555804729462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,2048,0.005643555687533484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,3072,0.012457778056462606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,1536,0.0041982221106688184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,2560,0.012827555338541666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,1024,0.003928889003064898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,2048,0.01279022213485506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,768,0.00351555562681622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,1536,0.011443555355072021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,512,0.0032417778339650896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,1024,0.011775111158688864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,256,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,768,0.01144266708029641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,128,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,512,0.010699555277824402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,64,0.002532444480392668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,256,0.011101333631409539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,4,32,32,0.0025555555605226094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,128,0.011749332977665795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,64,0.009081777599122789
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,4,32,32,0.010099555883142684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,16384,0.3046177758110894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,16384,0.21918043825361463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,12288,0.23499910036722818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,16384,0.17775021659003365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,12288,0.15711466471354166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,10240,0.19868977864583334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,12288,0.13997777303059897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,10240,0.14021510548061794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,10240,0.11931910779741074
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,8192,0.16211822297838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,8192,0.15766578250461155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,7168,0.14302044444613987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,8192,0.10117955340279473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,7168,0.08887910842895508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,7168,0.13600444793701172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,6144,0.1250391138924493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,6144,0.10500088665220474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,5120,0.10522488753000896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,5120,0.08457599745856391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,5120,0.06644977463616265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,4096,0.0896151132053799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,4096,0.07956088913811578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,6144,0.07742310894860162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,3584,0.06950666507085164
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,3584,0.07181066936916776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,3072,0.060416890515221484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,3584,0.04913155568970574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,3072,0.068067557281918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,3072,0.04266044497489929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,2560,0.05103733473353916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,2560,0.05095200075043572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,2560,0.03720888826582167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,4096,0.05308533377117581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,2048,0.045443554719289146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,1536,0.03259288933542039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,2048,0.0497573349210951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,2048,0.03109333250257704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,1024,0.02257866660753886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,1536,0.024753777517212763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,1536,0.03629866573545668
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,768,0.0184871107339859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,1024,0.028110222684012517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,1024,0.018611555298169453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,512,0.01331733332739936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,768,0.024088000257809956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,768,0.01594311164485084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,256,0.008784888519181145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,512,0.018386666973431904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,512,0.012878222597969903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,128,0.006309333360857434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,256,0.016606221596399944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,128,0.01348888874053955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,128,0.010110222631030613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,64,0.005623999983072281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,65536,32,0.005688000056478713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,64,0.01310488912794325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,65536,32,0.012809777425395118
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,65536,0.31615734100341797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,65536,0.16283733314938015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,16384,0.08562666840023464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,65536,256,0.009697777529557547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,16384,0.08425155613157485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,12288,0.06609600120120578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,16384,0.0499582224422031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,65536,0.1743617720074124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,10240,0.05650222301483154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,12288,0.06317866510815091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,12288,0.045347554816140064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,8192,0.04806222187148201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,10240,0.056721780035230845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,10240,0.030329777134789362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,7168,0.0358862214618259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,8192,0.04685599936379326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,8192,0.03285333183076646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,6144,0.031889776388804116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,7168,0.04313688808017307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,7168,0.03351644343800015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,5120,0.027140445179409448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,6144,0.03999555442068312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,4096,0.022502221994929846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,6144,0.02225422196918064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,5120,0.03296000096533034
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,5120,0.02390044430891673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,3584,0.02036622166633606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,4096,0.01927288870016734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,4096,0.028619554307725694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,3072,0.017609778377744887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,3584,0.027927110592524212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,3584,0.019819556011093985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,2560,0.015669332610236276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,3072,0.017111111018392775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,3072,0.025573333104451496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,2048,0.01551999979548984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,2560,0.015224888920783997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,1536,0.010728889041476779
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,2048,0.02095288866096073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,2048,0.013171555267439948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,1024,0.008422222402360704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,1536,0.01867111192809211
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,1536,0.010795555180973478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,1024,0.015848888291252982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,1024,0.009350222018029954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,768,0.007093333535724216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,768,0.015020444989204407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,768,0.00830577810605367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,512,0.004561777744028303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,512,0.007283555964628856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,256,0.003176888864901331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,2560,0.024753777517212763
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,256,0.012576888832781049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,128,0.0031582222630580268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,256,0.006271111054552927
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,16384,128,0.006536888993448681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,128,0.011069333387745751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,64,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,16384,32,0.0028355556229750314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,64,0.010757333702511258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,32,0.009405333134863112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,65536,0.22549332512749565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,16384,512,0.012836444709036084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,65536,0.14344622029198542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,16384,0.06401511033376057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,16384,0.040362666050593056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,16384,0.0668755571047465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,65536,0.15037777688768175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,12288,0.05053777827156914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,10240,0.04355555441644457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,12288,0.05151822169621786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,12288,0.03696444299485948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,8192,0.037535998556349016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,10240,0.04881599876615736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,10240,0.03453600075509813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,7168,0.032660444577534996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,8192,0.04682666725582547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,8192,0.022515555222829182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,6144,0.029940442906485662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,7168,0.04348266786999173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,7168,0.025277333127127752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,5120,0.025386666258176167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,6144,0.033287998702791
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,6144,0.024803555674023096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,4096,0.0221368885702557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,5120,0.02797777785195245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,4096,0.02643199927277035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,3584,0.019696000549528334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,3584,0.02630399995379978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,3584,0.017064889272054035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,3072,0.01737866633468204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,3072,0.024373332659403484
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,5120,0.0185208883550432
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,2560,0.01564711166752709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,3072,0.015617777903874716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,4096,0.016858667135238647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,2048,0.012025777664449481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,2560,0.013577777478430005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,2560,0.020510221521059673
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,1536,0.008683555656009251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,2048,0.011903111305501727
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,1536,0.01717688971095615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,1536,0.010052444206343757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,1024,0.014129777749379476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,1024,0.008471999731328752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,768,0.005807999935415056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,768,0.01349866638580958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,768,0.007687999970383114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,2048,0.019677332705921598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,512,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,1024,0.006984000404675801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,512,0.013088000317414602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,512,0.006934222247865465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,256,0.0035013332962989807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,128,0.0032400000426504347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,256,0.006281777802440856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,256,0.012459555433856117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,64,0.0028897778441508612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,128,0.010753778119881948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,12288,32,0.0029164445069101122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,12288,128,0.006266666783226862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,64,0.01036800030204985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,12288,32,0.010090666512648264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,65536,0.18991643852657744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,16384,0.055658664968278676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,65536,0.13860355483161077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,16384,0.05824177794986301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,65536,0.1421440045038859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,12288,0.04286844531695048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,12288,0.04561511013242933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,12288,0.03317688902219137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,10240,0.03752977649370829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,10240,0.04071911176045736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,10240,0.02535199953450097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,8192,0.030955556366178725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,8192,0.03459111187193129
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,8192,0.026319111386934917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,7168,0.03339644604259067
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,16384,0.04235644472969902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,6144,0.025817778375413682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,7168,0.023944889505704243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,5120,0.021848888860808477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,6144,0.020598222812016804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,6144,0.028370665179358587
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,5120,0.025591111845440332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,4096,0.019107555349667866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,5120,0.01886044442653656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,4096,0.022991999983787537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,3584,0.01752444439464145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,7168,0.028423110644022625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,3584,0.014020444618331062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,3584,0.02239199976126353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,3072,0.01648000048266517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,3072,0.0206675562593672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,3072,0.01349688900841607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,2560,0.011988444460762871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,2560,0.011659555964999728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,2560,0.018719111879666645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,2048,0.011418666276666852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,2048,0.010264889233642155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,2048,0.01717155509524875
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,1536,0.008633777499198914
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,4096,0.016008888681729633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,1024,0.007302222152551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,1536,0.015525332755512662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,1536,0.009280000295903947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,768,0.005666666560702854
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,1024,0.014131555954615274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,1024,0.007668444679843054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,512,0.004866666677925321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,768,0.012812444733248817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,768,0.007305777735180325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,256,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,512,0.012123555772834353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,512,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,128,0.0031964443624019623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,256,0.006640000061856375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,256,0.012440889245933957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,64,0.002970666728085942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,128,0.009680888719028896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,10240,32,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,10240,128,0.006486222147941589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,64,0.01069866700304879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,10240,32,0.008992888861232335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,65536,0.1553084452946981
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,16384,0.04561866654290093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,65536,0.08929955297046238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,65536,0.07670399877760145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,16384,0.047649777597851224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,12288,0.03685777717166477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,16384,0.024660444921917383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,10240,0.029704888661702473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,12288,0.03772799836264716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,12288,0.019329778022236295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,8192,0.023873777853118047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,10240,0.032250665956073336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,10240,0.016555555992656283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,7168,0.020744888318909537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,8192,0.02824977702564663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,8192,0.01587466730011834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,6144,0.018924444913864136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,7168,0.026350221700138513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,7168,0.015118221441904703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,5120,0.01608088943693373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,6144,0.02422399984465705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,6144,0.01312711089849472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,4096,0.014535999960369535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,5120,0.021524444222450256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,5120,0.012166221936543783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,3584,0.012664000193277994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,4096,0.02034311162100898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,4096,0.010846222440401712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,3072,0.010980444649855295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,3584,0.019392000304328073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,3584,0.01050222251150343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,2560,0.009966221948464712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,3072,0.018210666047202218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,3072,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,2048,0.009363555245929295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,2560,0.016943999462657504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,2560,0.00870666652917862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,1536,0.0069520001610120135
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,2048,0.015849777393870883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,2048,0.007993777592976889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,1024,0.004983110974232356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,1536,0.013792888985739814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,1536,0.007370666497283512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,768,0.004211555752489301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,1024,0.013046222428480784
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,1024,0.006592000110281839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,512,0.003705777641799715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,768,0.012851555314328937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,768,0.006333333336644703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,256,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,512,0.01105955574247572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,512,0.006243555496136348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,128,0.0032088889016045462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,256,0.011399110986126794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,256,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,64,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,128,0.011085333095656501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,8192,128,0.005602666487296422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,8192,32,0.0029075555503368378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,64,0.008993777963850234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,8192,32,0.009673777553770278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,65536,0.13491643799675837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,65536,0.0815111133787367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,16384,0.0421013335386912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,65536,0.07335466808742948
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,16384,0.04563200142648485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,16384,0.022535110513369244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,12288,0.032540443870756365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,12288,0.034318221939934626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,12288,0.01791111131509145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,10240,0.027794665760464136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,10240,0.03195022212134467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,10240,0.01593244406912062
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,8192,0.023852444357342188
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,8192,0.02757599949836731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,8192,0.014736889137162102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,7168,0.021215111017227173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,7168,0.025431111454963684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,7168,0.012787555654843649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,6144,0.017677333619859483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,6144,0.024284443921513025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,5120,0.014184888866212634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,6144,0.012109333442317115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,5120,0.011072888970375061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,4096,0.013464888764752282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,4096,0.020243555307388306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,4096,0.009547555612193214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,3584,0.011557333171367645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,3584,0.01891644464598762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,3584,0.009276444713274637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,3072,0.010401777923107147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,5120,0.020987555384635925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,2560,0.008985777695973715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,3072,0.008644444247086843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,3072,0.017768889665603638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,2560,0.016561778055297006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,2048,0.009319110876984065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,2560,0.007700444095664554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,1536,0.006640888750553131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,2048,0.014455111490355598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,2048,0.007527111305130853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,1024,0.004592888885074192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,1536,0.013784888717863293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,1536,0.006683555742104848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,768,0.004232888834344016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,1024,0.006284444282452266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,1024,0.01309333327743742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,512,0.0035662220584021676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,768,0.012068444656001197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,768,0.006315555423498154
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,256,0.003167111012670729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,256,0.011420444481902652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,128,0.0028213332924577924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,128,0.010054222411579555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,128,0.005591111050711737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,64,0.0029048888633648553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,64,0.009463110731707679
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,7168,32,0.0029048888633648553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,512,0.01185066666867998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,7168,32,0.009354666703277165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,65536,0.11743911107381184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,7168,256,0.005592000153329637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,16384,0.03624711102909512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,65536,0.07459466987186007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,16384,0.041928887367248535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,65536,0.07372977998521593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,12288,0.029273778200149536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,12288,0.0362097786532508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,12288,0.01718311177359687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,10240,0.025474665893448725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,10240,0.02922755479812622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,10240,0.016567111015319824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,8192,0.02128533356719547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,8192,0.02784088916248745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,8192,0.013839999834696451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,7168,0.0196142229768965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,7168,0.027592000034120347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,7168,0.013265777793195514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,6144,0.017664889494578045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,16384,0.022873777482244704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,5120,0.015505777464972602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,6144,0.02332266668478648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,6144,0.011728888584507836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,4096,0.012080888781282636
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,5120,0.020250666472646926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,5120,0.011033777561452655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,3584,0.010407999985747868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,4096,0.01924800044960446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,4096,0.009371555513805812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,3072,0.009386666946940953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,3584,0.018325333793958027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,3584,0.008302222523424361
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,2560,0.008359111017651027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,3072,0.017128000656763714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,3072,0.00870311094654931
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,2048,0.00814488861295912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,2560,0.016186666157510545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,2560,0.007617777420414819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,2048,0.014840000205569796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,2048,0.0069724445541699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,1536,0.006284444282452266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,1024,0.004590222405062782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,1536,0.01312622262371911
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,1024,0.006257777826653586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,768,0.0038542221817705366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,1024,0.01295377810796102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,768,0.01204622205760744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,512,0.0034684443639384378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,768,0.006267555471923616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,512,0.011063999599880643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,512,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,256,0.0032017777363459268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,256,0.005650666852792104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,256,0.011795555551846823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,128,0.0028444443725877335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,1536,0.0069306666652361555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,64,0.0027955555253558685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,6144,128,0.005572444448868434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,128,0.011038222246699862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,6144,32,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,64,0.010026666853162978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,65536,0.09990311331219143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,6144,32,0.010059555371602377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,16384,0.03204977843496535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,65536,0.06954666641023424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,65536,0.06834489107131958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,16384,0.034663110971450806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,12288,0.025440888272391424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,16384,0.021007999777793884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,10240,0.022464000516467627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,12288,0.016851555969980028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,12288,0.02958400050799052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,8192,0.019355555375417072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,10240,0.027594667341974046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,10240,0.015182221929232279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,7168,0.01719288859102461
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,8192,0.02465244465404087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,8192,0.013566222455766467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,6144,0.016414221790101793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,7168,0.01237422227859497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,7168,0.02614044480853611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,5120,0.014008000493049622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,6144,0.02200177808602651
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,6144,0.011075555450386472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,4096,0.011338666909270816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,5120,0.019513777560657926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,5120,0.009761778016885122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,3584,0.009356444080670675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,4096,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,4096,0.018346667289733887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,3072,0.00870400004916721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,3584,0.01760444376203749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,3584,0.008081778056091731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,2560,0.00796088907453749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,3072,0.0162791113058726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,3072,0.007907555335097844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,2048,0.007879111501905654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,2560,0.015076445208655464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,2560,0.007320000065697565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,1536,0.00544888898730278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,2048,0.0144533332851198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,2048,0.006940444310506185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,1024,0.003875555677546395
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,1536,0.013415111435784234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,1536,0.006608888920810487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,768,0.00388355553150177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,1024,0.012053333222866058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,1024,0.006292444550328785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,512,0.0032355555643637977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,768,0.005938666562239329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,768,0.012799111505349478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,256,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,512,0.01146755533085929
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,512,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,128,0.0028728888266616394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,256,0.011830222275522021
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,256,0.005683555371231503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,64,0.0029200000895394217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,128,0.009313777916961247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,5120,128,0.005647999958859549
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,5120,32,0.0028479999552170434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,64,0.01035911093155543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,65536,0.08104355467690362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,5120,32,0.009713778065310584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,16384,0.026991999811596338
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,65536,0.06419822242524889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,65536,0.06595466534296672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,16384,0.03250044584274292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,12288,0.021659556362364028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,16384,0.020184000333150227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,10240,0.018258665998776753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,12288,0.031013333135181006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,12288,0.016328889462682936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,8192,0.016550223032633465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,10240,0.025054223007626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,10240,0.01499644418557485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,7168,0.014754666222466363
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,8192,0.023846222294701472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,8192,0.012856889102194043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,7168,0.0236488895283805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,7168,0.011717333561844297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,6144,0.013711111413107978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,6144,0.00979377743270662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,5120,0.012215110990736218
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,5120,0.018572444717089336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,5120,0.009358222285906473
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,4096,0.012103111379676394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,4096,0.01754044493039449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,4096,0.008328888979223039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,3584,0.011414222419261932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,6144,0.02052622205681271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,3584,0.016181333197487723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,3584,0.007653333246707916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,3072,0.0101742222905159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,3072,0.015180443723996481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,3072,0.007592889169851939
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,2560,0.008700444466537898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,2560,0.014676445060306124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,2560,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,2048,0.007798222204049428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,2048,0.01385244478782018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,2048,0.006959111326270633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,1536,0.006573333508438534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,1536,0.013086222112178802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,1536,0.00665155549844106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,1024,0.004937777916590373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,1024,0.012412444584899478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,768,0.004912000149488449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,768,0.012113778127564324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,768,0.006308444258239534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,512,0.004366222355100843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,512,0.01145600030819575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,256,0.003198222153716617
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,512,0.0059057776298787855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,256,0.00563733321097162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,256,0.011752000285519494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,128,0.0028231110837724474
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,1024,0.006254222244024277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,64,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,128,0.0110871113008923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,4096,128,0.005631111148330901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,4096,32,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,64,0.009682666924264696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,65536,0.07401333252588908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,4096,32,0.009115555220180089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,16384,0.02353777819209629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,65536,0.0627066691716512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,65536,0.06544888681835599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,12288,0.019055111540688407
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,16384,0.032456000645955406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,16384,0.01982399986849891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,10240,0.016528889536857605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,12288,0.026575111680560645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,12288,0.016200888488027785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,8192,0.01445777714252472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,10240,0.02570755614174737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,10240,0.014649778604507446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,7168,0.013758222262064615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,8192,0.012437333663304647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,7168,0.02162577708562215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,7168,0.011047999891969891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,6144,0.012729778057999082
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,6144,0.020248000820477802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,6144,0.010227555202113258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,5120,0.011961778004964193
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,5120,0.01791200041770935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,5120,0.00903555585278405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,4096,0.010724444356229572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,4096,0.008068444828192392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,3584,0.010729778144094678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,8192,0.022755554980701868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,3584,0.015196444259749519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,3072,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,3584,0.007939555578761632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,3072,0.01461955573823717
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,3072,0.007497777541478475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,2560,0.008359111017651027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,2560,0.013412444127930535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,2560,0.007323555648326874
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,2048,0.007321777443091075
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,2048,0.01349866638580958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,2048,0.006978666616810693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,4096,0.016534222496880423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,1536,0.006248000181383557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,1024,0.004922666483455234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,1536,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,1536,0.012461333639091916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,768,0.004592888885074192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,1024,0.012372444073359171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,1024,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,512,0.0041911109454101985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,768,0.006248888870080312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,768,0.012158222496509552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,256,0.003141333245568805
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,512,0.011714666253990598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,512,0.005906666732496685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,128,0.0028906667398081887
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,256,0.011379555695586734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,256,0.005559999909665849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,128,0.011007111105653973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,64,0.0028568889117903183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,64,0.010382222632567087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3584,32,0.0029084444459941653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3584,32,0.010071111222108206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,65536,0.0637635588645935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,65536,0.06054044432110257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,16384,0.020950223008791607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3584,128,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,65536,0.06422399812274508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,12288,0.01829777823554145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,16384,0.01959822244114346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,12288,0.015989333391189575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,12288,0.026199110680156287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,10240,0.015221332510312399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,10240,0.014389332797792224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,10240,0.024007999234729346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,8192,0.013631111217869652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,8192,0.011703111231327057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,8192,0.021746666895018682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,7168,0.012636444634861417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,16384,0.028891556792789038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,7168,0.010643555886215635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,6144,0.012390221986505719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,7168,0.022631999519136217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,5120,0.010671111444632212
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,6144,0.01925066610177358
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,6144,0.009714666340086196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,4096,0.010400888820489248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,5120,0.016189333465364244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,5120,0.009006222089131674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,3584,0.009304888546466827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,4096,0.01518311103185018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,4096,0.008405333591832055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,3072,0.00866844422287411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,3584,0.014533332652515836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,3072,0.014146667387750415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,3072,0.007651555869314406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,2560,0.00795288880666097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,2560,0.013775111072593264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,2560,0.007296889192528195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,2048,0.006485333459244833
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,2048,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,2048,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,1536,0.006631999793979857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,1536,0.012489777472284106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,1536,0.006614222294754452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,1024,0.004909333255555895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,3584,0.007910221815109253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,768,0.0052560000783867305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,1024,0.011400000088744693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,768,0.005951111101441913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,768,0.011789333489206104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,512,0.004242666479614046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,512,0.005594666633341048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,256,0.00351555562681622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,256,0.010716444088353051
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,256,0.005657777604129579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,128,0.0032266666077905228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,1024,0.006173333360089197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,128,0.009353777600659264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,64,0.0028640000770489373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,3072,128,0.005597333527273602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,3072,32,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,64,0.008707555631796518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,512,0.011137777732478248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,3072,32,0.009344889058007134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,65536,0.056840889983707
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,16384,0.019426667028003268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,65536,0.05862489011552599
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,65536,0.05982933441797892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,16384,0.027591110931502447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,16384,0.01943733294804891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,12288,0.023711111810472276
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,12288,0.015280889140235053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,10240,0.013863110707865821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,10240,0.021957332889238994
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,8192,0.012248000337017907
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,8192,0.019909333851602342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,8192,0.011039999624093374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,7168,0.011398221883508893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,12288,0.016346666547987197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,7168,0.019832889238993328
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,7168,0.010630221830474006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,6144,0.01148622234662374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,6144,0.017822222577200997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,6144,0.009630222287442949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,5120,0.010144888526863521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,10240,0.013109332985348172
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,4096,0.009767110976907942
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,5120,0.015169777803950839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,5120,0.009012444151772393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,3584,0.008903111020723978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,4096,0.014517333772447375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,4096,0.008049777812427944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,3072,0.007727999654081132
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,3584,0.014197332991494073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,3584,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,2560,0.007334222396214803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,3072,0.013813333378897773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,3072,0.007625777688291337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,2560,0.013390222357379066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,2560,0.007280000381999546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,2048,0.005975111077229182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,2048,0.013105777402718862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,1536,0.005942222144868638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,1536,0.012416000167528788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,1024,0.0048942222363419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,1536,0.006752000086837345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,1024,0.012120000190205045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,768,0.004770666774776247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,1024,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,768,0.01236177815331353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,512,0.0038746665749284956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,768,0.00600622221827507
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,2048,0.0069288888739215
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,512,0.005623999983072281
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,256,0.00360000009338061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,256,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,256,0.011476444701353708
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,128,0.003237333355678452
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2560,128,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,128,0.010760889285140567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,64,0.0028826666788922418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2560,32,0.0028275555620590844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,64,0.010029333333174387
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,32,0.0099982221921285
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,65536,0.04794666502210829
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2560,512,0.011402666568756104
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,65536,0.057068447271982826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,16384,0.01651822196112739
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,65536,0.05873777468999227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,12288,0.013340444200568728
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,16384,0.024865777956114873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,16384,0.01882933411333296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,10240,0.01201777739657296
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,12288,0.02180800007449256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,12288,0.01479555500878228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,8192,0.010991999672518836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,10240,0.02032888929049174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,10240,0.012099555797047086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,7168,0.010410666465759277
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,8192,0.011023110813564725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,8192,0.018280888597170513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,6144,0.010195555786291758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,7168,0.01604622271325853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,7168,0.010363555616802638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,5120,0.008994667066468133
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,6144,0.015867556134859722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,6144,0.009649777577983009
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,4096,0.00977955593003167
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,5120,0.014428445034556918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,5120,0.00904177791542477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,3584,0.008964444200197855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,4096,0.008343111309740279
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,4096,0.013870221873124441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,3072,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,3584,0.013472889032628802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,3584,0.007986666427718269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,2560,0.005903111149867375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,3072,0.01331999980741077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,3072,0.007611555357774098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,2048,0.006715555571847492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,2560,0.013089777694808112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,2560,0.007187555233637492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,1536,0.006007110906971826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,2048,0.012727110750145383
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,2048,0.0069191112286514705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,1024,0.004593777573770947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,1536,0.012416000167528788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,1536,0.006672888994216919
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,768,0.004219555606444676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,1024,0.012093333734406365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,1024,0.006270222365856171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,512,0.003920000046491623
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,768,0.011725333001878528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,768,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,256,0.003510222252872255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,512,0.005753777921199799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,256,0.011757333245542316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,256,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,128,0.0031840000301599503
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,128,0.011034666664070554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,2048,128,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,64,0.0029226665695508323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,2048,32,0.0028000000036425064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,64,0.009671111073758867
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,65536,0.03644888930850559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,65536,0.055870221720801465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,512,0.011155555645624796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,16384,0.013833777772055732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,65536,0.05827555391523573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,16384,0.024117334021462336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,16384,0.01683911184469859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,12288,0.01144444445768992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,12288,0.013479111095269522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,10240,0.013549333645237817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,2048,32,0.010084444450007545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,10240,0.018775999546051025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,10240,0.012392000191741519
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,8192,0.012083555261294046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,8192,0.0170151111152437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,8192,0.01074488874938753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,7168,0.01143022212717268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,12288,0.021499555971887376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,7168,0.015606222881211175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,6144,0.010491555763615502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,7168,0.010424888796276517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,6144,0.01516711049609714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,5120,0.00944266633854972
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,5120,0.014548444085650973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,5120,0.008991999758614434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,4096,0.008651555412345463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,4096,0.008002666963471307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,3584,0.0069884442620807225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,4096,0.013547555440002017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,3584,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,3072,0.006301333506902059
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,3584,0.008079110748238033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,6144,0.009763555394278632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,2560,0.005924444645643234
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,3072,0.013386666774749756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,3072,0.007708444363541073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,2048,0.006677333265542984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,2560,0.01312444441848331
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,2560,0.007293333609898885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,1536,0.0052639999323421055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,2048,0.006959999601046245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,2048,0.011760888828171624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,1024,0.004942222187916438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,1536,0.012400888734393649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,1536,0.006728000111050076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,768,0.0038728887836138406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,1024,0.01109777804878023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,1024,0.006256888724035687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,512,0.0038577777643998465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,768,0.012063110868136087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,768,0.00591111100382275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,256,0.0032195556494924757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,512,0.011072888970375061
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,512,0.005574222240183089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,128,0.0031831111345026228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,256,0.005624889085690181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,128,0.010743111371994019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,64,0.0028435554769304064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1536,128,0.005587555468082428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1536,32,0.0029093333416514923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,64,0.009676444033781687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,65536,0.026568889617919922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,65536,0.0528657767507765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,65536,0.05787377887301975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,16384,0.014493332968817817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,256,0.01074222226937612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,16384,0.022300443715519373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,16384,0.015934222274356417
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,12288,0.012307555311255984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,12288,0.013804444836245643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,12288,0.01889244384235806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1536,32,0.010116444693671333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,10240,0.011055111057228513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,8192,0.009656888743241629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,10240,0.01742488808102078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,8192,0.01624000072479248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,7168,0.009129777550697327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,8192,0.010709332923094431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,7168,0.015885333220163982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,7168,0.010059555371602377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,6144,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,6144,0.014874666929244995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,5120,0.006686222222116258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,5120,0.01388088862101237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,5120,0.009029332962301042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,4096,0.007393777370452881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,4096,0.013445333474212222
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,4096,0.008013333711359236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,3584,0.006969778074158563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,6144,0.009656888743241629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,3584,0.013196444345845116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,3584,0.007977777885066139
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,3072,0.013426666458447775
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,10240,0.012387555506494312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,2560,0.005625777774386936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,3072,0.007633777956167857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,2560,0.013095999757448832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,2048,0.006370666540331311
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,2560,0.007289778027269576
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,2048,0.0069679998689227635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,2048,0.01274222218328052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,1536,0.005551110953092575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,1536,0.012110222544935016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,3072,0.006319111006127463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,1024,0.0048942222363419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,1024,0.006333333336644703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,1024,0.011047110789351992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,768,0.004705777598751916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,768,0.011429333024554782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,768,0.005994666781690385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,512,0.003951111187537511
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,512,0.01111288865407308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,512,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,256,0.0032462223122517266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,256,0.010732444624106089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,128,0.003185777821474605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,1536,0.006618666566080517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,256,0.00563733321097162
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,64,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,1024,128,0.005596444424655702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,128,0.009387555221716562
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,1024,32,0.0028968888024489083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,65536,0.02096533278624217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,64,0.00905600024594201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,1024,32,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,16384,0.012850667039553324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,65536,0.05240355597601997
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,65536,0.057229333453708224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,12288,0.010450666977299584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,16384,0.020607110526826646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,16384,0.016507556041081745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,10240,0.009328888522254096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,12288,0.018244443668259513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,12288,0.013444444371594323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,8192,0.008655110994974772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,10240,0.01754044493039449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,10240,0.012452444268597497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,7168,0.00833511104186376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,8192,0.015826667348543804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,8192,0.010954666468832227
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,6144,0.007954667011896769
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,7168,0.015511110424995422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,6144,0.015198222464985318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,6144,0.009728888670603434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,5120,0.014103111293580798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,5120,0.009089777866999308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,4096,0.007316444483068254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,4096,0.013759999639458127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,7168,0.010053333308961656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,4096,0.008294222255547842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,3584,0.007301333049933116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,5120,0.006715555571847492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,3584,0.013413333230548434
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,3072,0.006621333460013072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,3072,0.007631999750932057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,2560,0.006258666515350342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,2560,0.013129777378506131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,2560,0.007338666253619724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,2048,0.005285333428117964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,2048,0.012503999802801343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,3584,0.007680888805124495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,2048,0.006948444578382704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,3072,0.013622221847375235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,1536,0.006772444479995304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,1024,0.00479644454187817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,1024,0.011739555332395764
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,1024,0.00629688882165485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,768,0.003914666672547658
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,1536,0.005586666779385672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,768,0.011782222323947482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,512,0.0035146665241983202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,512,0.011405333048767514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,512,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,256,0.003256888853179084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,256,0.01108977778090371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,256,0.005603555589914322
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,128,0.0031511110977994073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,768,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,128,0.010026666853162978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,768,128,0.005292444593376583
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,64,0.008363555702898238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,32,0.002899555489420891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,32,0.0083155557513237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,65536,0.022707555029127333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,65536,0.05105688836839464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,768,1536,0.012065778175989786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,16384,0.010007999837398529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,65536,0.05401510993639628
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,768,64,0.0029128889242808023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,12288,0.008444444172912175
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,16384,0.021006221572558086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,16384,0.01593155496650272
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,10240,0.007612444460391998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,12288,0.01824711097611321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,12288,0.01369599997997284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,8192,0.008079110748238033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,10240,0.016851555969980028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,10240,0.012088889049159156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,7168,0.007657777931955125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,8192,0.016151999433835346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,8192,0.010500444306267632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,6144,0.007299555672539606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,7168,0.015610666738616096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,7168,0.010315555665228102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,5120,0.006955555743641323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,6144,0.014549333188268872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,6144,0.0094035557574696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,4096,0.007137777904669444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,5120,0.014088888963063559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,5120,0.008685333033402761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,4096,0.01351999988158544
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,4096,0.00816000004609426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,3584,0.013095999757448832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,3584,0.007981333467695449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,3072,0.01347733371787601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,3072,0.007622222105662028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,2560,0.006250666661394968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,2560,0.01309866706530253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,3584,0.006953777538405524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,2560,0.007251555720965068
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,3072,0.006606222026877933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,2048,0.012472000386979846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,1536,0.004592888885074192
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,2048,0.006626666833957036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,1536,0.01204177737236023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,1024,0.004256000121434529
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,1536,0.006252444452709622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,1024,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,1024,0.005639111002286275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,768,0.003907555507289039
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,768,0.011405333048767514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,768,0.005615111026499007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,2048,0.005904888941182031
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,512,0.0035439998739295532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,256,0.003024000053604444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,512,0.011121778024567498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,256,0.01072177787621816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,128,0.0028648889727062653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,256,0.005478222337034013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,128,0.009364444348547194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,128,0.005557333429654439
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,64,0.009047110875447592
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,32,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,512,32,0.009000000026490953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,65536,0.014488889111412896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,512,512,0.005568888866239124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,65536,0.0473582214779324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,16384,0.00904088881280687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,65536,0.049351112710105054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,512,64,0.0028417776856157514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,16384,0.016574222180578444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,16384,0.0206631107462777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,12288,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,10240,0.007628444168302748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,12288,0.013491555220550962
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,10240,0.01683644453684489
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,8192,0.006976889239417181
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,10240,0.012125333150227865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,8192,0.010732444624106089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,8192,0.015572445260153877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,7168,0.006665777828958299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,7168,0.015234667393896313
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,7168,0.010064889159467485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,6144,0.006623999940024481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,6144,0.01482311056719886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,5120,0.006305777778228124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,12288,0.018534221582942538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,5120,0.013815110756291283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,4096,0.005917333480384614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,5120,0.008967999782827165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,4096,0.013424889081054263
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,4096,0.008200888832410177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,3584,0.006272000157170826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,3584,0.007670222057236566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,3584,0.013450666434235044
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,3072,0.005951111101441913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,3072,0.013097777962684631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,6144,0.009359999663299983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,2560,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,2560,0.01221422188811832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,2048,0.005930666708283954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,2048,0.011406222151385413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,2048,0.006613333192136552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,1536,0.004550222307443619
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,1536,0.011111111276679568
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,3072,0.007600888609886169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,1024,0.003961777935425441
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,1536,0.006269333263238271
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,2560,0.007183111376232571
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,1024,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,768,0.003607999947335985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,768,0.005946666830115848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,768,0.011805333197116852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,512,0.003547555456558863
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,512,0.011391999820868174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,512,0.005587555468082428
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,256,0.002893333426780171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,256,0.011744000017642975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,256,0.00553688903649648
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,128,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,128,0.010056888891590966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,128,0.005604444278611078
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,64,0.002534222271707323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,256,32,0.0026444445053736367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,256,1024,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,64,0.009350222018029954
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,65536,0.01346933344999949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,256,32,0.009031111167536842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,16384,0.007657777931955125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,65536,0.0473235547542572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,65536,0.04763199885686239
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,12288,0.007360888852013483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,16384,0.01622222198380364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,16384,0.020690666304694284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,12288,0.017893332574102614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,10240,0.007106666763623555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,12288,0.013402666482660504
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,8192,0.006599999964237213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,10240,0.01684266659948561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,10240,0.012268444730175866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,8192,0.015861334072219003
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,8192,0.010475555227862464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,7168,0.015476443701320224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,7168,0.010017777482668558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,6144,0.006233777850866318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,6144,0.014191110928853353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,6144,0.009705777797434065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,5120,0.005961777849329843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,5120,0.0087333329849773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,4096,0.005598222215970357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,4096,0.013085333009560903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,7168,0.006686222222116258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,3584,0.005952888892756567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,4096,0.00795911086930169
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,3584,0.007672888537247975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,3584,0.01346577786737018
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,3072,0.0059164443777667145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,3072,0.013071111506885953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,3072,0.007272889216740926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,2560,0.00591822216908137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,5120,0.013770666387346057
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,2560,0.011748444702890186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,2048,0.005314666777849197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,2560,0.007352888584136963
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,1536,0.004376888689067629
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,2048,0.011374221907721626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,2048,0.006903999795516332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,1024,0.0038613333470291565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,1536,0.012171555724408893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,1536,0.006244444598754247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,768,0.0035422220826148987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,1024,0.011750222080283694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,1024,0.006304889089531369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,512,0.0032391111469931076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,768,0.005752000129885144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,512,0.011169777976142036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,256,0.002874666617976295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,512,0.005605333381228977
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,256,0.00555377784702513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,128,0.0028168888141711554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,128,0.010036444498433007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,2,128,128,0.0052968888647026485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,64,0.0028817777832349143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,768,0.010867555936177572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,64,0.009392889009581672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,128,32,0.0028204443968004654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,65536,0.009715555442704095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,32,0.009101333717505137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,16384,0.006627555522653792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,65536,0.04625244604216682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,128,256,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,12288,0.006311999840868845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,10240,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,16384,0.020592888196309406
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,8192,0.005962666538026597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,12288,0.018238221605618794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,10240,0.016521778371598985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,8192,0.01590044465329912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,6144,0.005960000058015187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,7168,0.01512977812025282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,5120,0.005935110979610019
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,6144,0.01479466590616438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,4096,0.005420444326268301
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,5120,0.013434666726324292
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,3584,0.005552000055710475
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,4096,0.013519110778967539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,3072,0.005616888817813661
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,3584,0.012808000048001608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,2560,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,3072,0.012875555290116204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,2048,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,2560,0.011731555892361535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,7168,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,1536,0.004265777766704559
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,1024,0.003896000070704354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,2048,0.012091555529170565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,768,0.003543111185232798
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,1536,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,1024,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,768,0.011704000333944956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,256,0.0029146667155954572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,512,0.01038577738735411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,128,0.0028524444335036804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,256,0.011074666347768573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,64,0.0028373334142896864
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,128,0.009167110754383935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,32,0.0028355556229750314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,64,0.009030222064918941
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,65536,0.009010666774378883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,64,32,0.0083155557513237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,16384,0.006285333385070165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,65536,0.046557333734300405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,16384,0.020000000794728596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,64,512,0.003206222214632564
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,12288,0.018232888645595975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,10240,0.005951111101441913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,10240,0.017064000169436138
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,8192,0.005344888816277186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,8192,0.01587199999226464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,7168,0.005960000058015187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,7168,0.015411555767059326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,12288,0.006259555617968242
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,6144,0.005943111247486538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,5120,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,6144,0.014159111513031853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,4096,0.0052328887912962176
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,5120,0.014123555686738757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,3584,0.005884444548024072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,4096,0.01312888910373052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,3072,0.005249777601824866
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,3584,0.013463111387358772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,2560,0.006237333433495627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,3072,0.01219555570019616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,2048,0.005559999909665849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,2560,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,1536,0.004557333472702238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,2048,0.011409777734014722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,1024,0.003903999924659729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,1536,0.012131555212868584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,768,0.003559999995761447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,512,0.003286222202910317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,1024,0.010855110983053843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,256,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,768,0.011709333293967776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,128,0.002847111059559716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,512,0.011377777490350934
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,64,0.0028355556229750314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,256,0.010046222143703038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,2,32,32,0.002871111035346985
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,128,0.011023110813564725
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,64,0.009379555781682333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,2,32,32,0.008025777836640676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,16384,0.3046026759677463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,16384,0.25198043717278373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,12288,0.23292711046006945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,16384,0.1981546613905165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,12288,0.2004284461339315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,10240,0.19674489233228895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,12288,0.14875377549065485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,10240,0.1852888928519355
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,8192,0.16144711441463894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,10240,0.12378756205240886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,8192,0.1488142278459337
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,7168,0.1412284506691827
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,8192,0.10047466887368096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,7168,0.13091911209954157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,6144,0.12424622641669379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,7168,0.09091022279527451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,6144,0.10740355650583903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,5120,0.10477688577440049
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,6144,0.07736266983879937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,5120,0.09830133120218913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,4096,0.08798133002387153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,5120,0.06716889142990112
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,3584,0.07013777891794841
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,4096,0.0830577810605367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,3584,0.07107555866241455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,3072,0.061443554030524365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,3584,0.049061334795422025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,3072,0.06179022126727634
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,3072,0.04310133390956455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,2560,0.05214222272237142
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,2560,0.0366888874106937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,4096,0.053265776899125844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,2048,0.04551466637187534
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,2048,0.04560177856021457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,2048,0.031104889180925157
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,1536,0.032856888241238065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,2560,0.05216000146336026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,1536,0.036438223388459944
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,1536,0.02382844520939721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,1024,0.022872888379626807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,768,0.018237334158685472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,1024,0.018431110514534842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,768,0.023287110858493384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,512,0.013091555900043912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,768,0.015265777707099915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,512,0.018231110440360177
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,512,0.012689777546458773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,256,0.008357333640257517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,256,0.009700444837411245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,128,0.0062773335311147906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,128,0.011723555624485016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,65536,128,0.009733333355850643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,64,0.005628444254398346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,64,0.011711999773979187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,65536,32,0.005589333259397083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,32,0.011433777709801992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,256,0.015656888484954834
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,65536,0.3153066635131836
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,16384,0.08564888768725926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,65536,0.27476710743374294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,65536,1024,0.026644445127911035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,65536,0.1838222212261624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,12288,0.06629511382844713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,16384,0.07449777921040852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,16384,0.06276355849372016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,10240,0.056808001465267606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,12288,0.06549421946207683
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,12288,0.05487733417087131
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,8192,0.04752444558673435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,10240,0.05862310859892103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,10240,0.04220622115665012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,7168,0.03816711240344577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,8192,0.03375022278891669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,7168,0.04779111014472114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,7168,0.031553778383466936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,6144,0.03277600142690871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,6144,0.04048444496260749
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,6144,0.028213332096735638
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,5120,0.028143111202451918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,5120,0.03447644578086005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,5120,0.024643555283546448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,4096,0.024033778243594702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,4096,0.03160355488459269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,4096,0.020641777250501845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,3584,0.020263999700546265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,3584,0.02815822097990248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,3584,0.01907199952337477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,3072,0.018294221825069852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,3072,0.025402666793929204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,2560,0.015601777368121676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,2560,0.023365333676338196
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,2560,0.01534222231970893
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,8192,0.05306666758325365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,2048,0.014189332723617554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,2048,0.013232000172138214
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,2048,0.020258666740523446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,1536,0.011043555206722684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,1536,0.018217777212460835
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,1536,0.011040888726711273
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,1024,0.00830844458606508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,3072,0.01701866587003072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,768,0.00700444479783376
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,1024,0.015853333804342482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,1024,0.009327111144860586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,512,0.00462666650613149
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,768,0.015153777268197803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,768,0.007972444097201029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,256,0.0035377778112888336
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,512,0.012418666647540199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,512,0.007287999822033777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,128,0.003236444460021125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,256,0.011716444459226398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,256,0.0063279999627007385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,64,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,128,0.010791999598344168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,16384,32,0.0032382222513357797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,16384,128,0.006605333338181178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,64,0.009335999687512716
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,16384,32,0.010746666954623329
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,65536,0.2245244450039334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,16384,0.06405599912007649
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,65536,0.19085955619812012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,65536,0.155114663971795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,16384,0.055959999561309814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,16384,0.07262221972147624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,12288,0.05047733253902859
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,10240,0.04342755675315857
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,12288,0.059243553214603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,12288,0.039480888181262545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,8192,0.03673422336578369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,10240,0.04878222280078464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,10240,0.03804888990190294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,7168,0.03221599923239814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,8192,0.045838223563300244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,8192,0.028452446063359577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,6144,0.02934755550490485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,7168,0.04337777694066366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,7168,0.025616000096003216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,5120,0.025813332862324182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,6144,0.03432622220781114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,6144,0.025019556283950806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,4096,0.021936888496081035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,5120,0.03168977631462945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,5120,0.020767110917303298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,3584,0.02001688877741496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,4096,0.02840355700916714
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,4096,0.01739022301303016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,3072,0.017519111434618633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,3584,0.0258559998538759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,3584,0.017186666528383892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,2560,0.015584889385435315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,3072,0.023678221636348303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,3072,0.01484444406297472
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,2048,0.012258666257063547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,2560,0.021333333518770006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,2560,0.013056000073750814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,1536,0.008799110849698385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,2048,0.019356444478034973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,2048,0.011743110915025076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,1024,0.006999111009968652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,1536,0.01795377830664317
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,1536,0.009879111415810054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,1024,0.014783110883500842
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,1024,0.008354666332403818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,768,0.005940444353553984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,768,0.014520888527234396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,768,0.007650666766696506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,512,0.003965333518054751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,512,0.013157332936922709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,512,0.0070168889231152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,256,0.0032044444233179092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,256,0.011745777395036487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,128,0.0031617778456873367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,128,0.01071111112833023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,128,0.006461333483457565
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,64,0.0028364445186323593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,64,0.010402667025725046
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,12288,32,0.0032471112079090546
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,12288,32,0.010454222559928894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,12288,256,0.006613333192136552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,65536,0.18909777535332573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,16384,0.05532800157864889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,65536,0.16984977987077501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,65536,0.1506177716785007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,12288,0.04242755638228523
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,16384,0.04667644368277656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,16384,0.05683022075229221
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,10240,0.037435554795795016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,12288,0.04749066630999247
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,12288,0.0361697773138682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,8192,0.030489779180950586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,10240,0.02976177798377143
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,8192,0.03432888785998026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,7168,0.028212444649802312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,8192,0.025035555164019268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,7168,0.03207111027505662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,6144,0.026105778084860906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,6144,0.028351111544503108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,6144,0.021351110604074266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,5120,0.022077333596017625
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,10240,0.04123911261558533
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,5120,0.025507556067572698
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,5120,0.01790488925245073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,4096,0.019469334019554984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,7168,0.023021333747439917
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,4096,0.0162124451663759
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,3584,0.01796444422668881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,3584,0.02216800053914388
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,3584,0.014740443891949125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,3072,0.016759999924235873
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,3072,0.01313511116637124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,2560,0.012246222131782107
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,2560,0.01918577816751268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,2560,0.01144444445768992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,2048,0.01184088902340995
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,4096,0.022589332527584497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,2048,0.01772799922360314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,2048,0.009685333404276106
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,1536,0.01609511176745097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,3072,0.02068800065252516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,1024,0.007117333511511485
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,1536,0.008660444782839881
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,1024,0.013748444616794586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,768,0.005912888795137405
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,1024,0.007675555845101674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,768,0.007103111180994246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,512,0.004884444591071871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,512,0.012725333372751871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,512,0.0069164443347189165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,256,0.003165333428316646
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,256,0.010821333361996545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,256,0.006287999865081575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,128,0.0032186667538351486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,768,0.013412444127930535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,128,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,64,0.0029111111329661477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,10240,128,0.00654666663871871
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,32,0.0028400001012616684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,64,0.009734222458468543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,10240,32,0.010021333065297868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,65536,0.15392533938090006
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,65536,0.1333813269933065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,16384,0.04569066564242045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,65536,0.07928444279564752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,16384,0.04706577791108025
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,16384,0.025397333833906386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,12288,0.03715733355946011
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,10240,1536,0.009095999929640029
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,10240,0.028534223635991413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,12288,0.037427554527918495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,12288,0.021993777818149988
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,8192,0.023629332582155865
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,10240,0.032935112714767456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,7168,0.022387555903858606
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,8192,0.02933599882655674
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,10240,0.018983110785484314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,6144,0.018645332919226754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,7168,0.026901332868470088
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,7168,0.0158906661801868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,5120,0.015834665960735746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,6144,0.013989332649442883
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,6144,0.025136889682875738
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,5120,0.022314666046036616
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,5120,0.01311822235584259
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,4096,0.014909333652920194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,4096,0.01108888867828581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,4096,0.020271110865804885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,3584,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,8192,0.016492444607946608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,3584,0.01013422260681788
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,3584,0.019015999303923715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,3072,0.011387555963463254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,2560,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,3072,0.017695999807781644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,3072,0.009369778136412302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,2560,0.017302221722073026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,2560,0.008685333033402761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,2048,0.015862221519152325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,1536,0.007306666837798224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,2048,0.008239111138714684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,1536,0.014144889182514615
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,1536,0.0074275558193524676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,1024,0.005274666680230035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,1024,0.012803555362754397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,1024,0.006629333313968446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,768,0.004495111190610462
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,768,0.012771555946932899
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,512,0.003942222230964237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,2048,0.009723555710580613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,512,0.012122666670216454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,256,0.0032319999817344877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,256,0.01069866700304879
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,256,0.005899555567238066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,128,0.0029208889851967492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,768,0.006584888945023219
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,128,0.00941244430012173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,128,0.00590844452381134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,64,0.0029066666546795103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,8192,512,0.005962666538026597
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,8192,32,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,64,0.01035377797153261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,65536,0.1351813342836168
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,8192,32,0.008645333349704742
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,16384,0.04072533382309808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,65536,0.10882933272255792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,16384,0.04178577661514282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,65536,0.07468266619576348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,16384,0.022474666436513264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,12288,0.032144887579811945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,10240,0.027664000789324444
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,12288,0.03571110963821411
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,12288,0.01795644395881229
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,8192,0.023696889479955036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,10240,0.032010667853885226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,10240,0.016359999775886536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,7168,0.02126844392882453
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,8192,0.02815022071202596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,8192,0.01428711083200243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,6144,0.01683466633160909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,7168,0.01346933344999949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,7168,0.02569688856601715
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,5120,0.015180443723996481
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,6144,0.012442666623327466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,6144,0.024882665938801233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,4096,0.012679999901188744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,5120,0.020747555626763236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,5120,0.011208888557222156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,3584,0.011206222077210745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,4096,0.009689778089523315
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,4096,0.020227554771635268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,3072,0.010027555955780877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,3584,0.009000889129108852
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,3584,0.0192266669538286
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,2560,0.009342222577995723
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,3072,0.01753155555990007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,3072,0.00833155545923445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,2048,0.00905866672595342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,2560,0.016918222109476726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,2560,0.007610666255156199
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,1536,0.006430222342411677
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,2048,0.007302222152551015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,2048,0.0147733340660731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,1024,0.004898666507667965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,1536,0.013578666581047906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,1536,0.006682666639486949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,768,0.004299555387761858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,1024,0.006312888943486744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,1024,0.01275644451379776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,768,0.013088888592190213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,512,0.003532444437344869
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,768,0.006067555397748947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,256,0.0032026666320032547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,512,0.012359999948077731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,512,0.005998222364319696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,128,0.002973333415057924
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,256,0.005919111271699269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,7168,128,0.005965333431959152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,128,0.00941777808798684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,64,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,7168,32,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,64,0.00980088859796524
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,256,0.009703111317422655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,7168,32,0.009494221872753566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,65536,0.11683200465308295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,65536,0.11682221624586318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,65536,0.07072355349858601
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,16384,0.036453333165910505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,16384,0.03879022267129686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,16384,0.02245866590076023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,12288,0.028985778490702312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,12288,0.017242666747834947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,12288,0.03204800022972955
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,10240,0.024957334001859028
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,10240,0.015563555889659457
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,8192,0.021070222059885662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,10240,0.027842667367723253
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,7168,0.019339554839664035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,8192,0.02887822190920512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,8192,0.014561777313550314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,6144,0.01794933279355367
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,7168,0.012423111332787408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,6144,0.011760888828171624
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,5120,0.015390222271283468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,6144,0.022172444396548804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,5120,0.022436444958051045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,5120,0.011164444188276926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,4096,0.011942221886581846
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,4096,0.00906844437122345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,3584,0.010208889014191097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,4096,0.018975110517607797
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,7168,0.02608977754910787
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,3072,0.009357333183288574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,3584,0.018597332967652213
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,3584,0.008791999684439765
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,2560,0.008866666919655269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,3072,0.007677333222495184
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,3072,0.01752800080511305
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,2048,0.008150222400824228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,2560,0.00737955586777793
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,2560,0.016515556308958266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,2048,0.014549333188268872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,1536,0.006361777583758037
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,2048,0.0069724445541699724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,1024,0.004262222184075249
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,1536,0.014177777700954013
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,1536,0.006958222223652734
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,1024,0.011785777906576792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,1024,0.006365333166387346
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,768,0.011393778026103973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,768,0.006107555495368109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,512,0.0035617777870761026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,512,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,512,0.012118221984969245
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,256,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,256,0.01073333372672399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,256,0.005701333284378052
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,128,0.002846222163902389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,128,0.008746667040718926
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,6144,128,0.00592533333433999
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,64,0.0028400001012616684
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,32,0.0029208889851967492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,64,0.009549332989586724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,6144,32,0.00867377801073922
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,65536,0.10147288772794937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,65536,0.08835022317038642
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,65536,0.06954577896330091
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,16384,0.03155644403563605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,16384,0.02088800072669983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,16384,0.03972711165746053
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,12288,0.02548888822396596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,6144,768,0.00397155558069547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,10240,0.02251200046804216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,12288,0.03166844447453817
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,12288,0.016890666551060148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,8192,0.019334221879641216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,10240,0.015172445111804538
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,10240,0.02591288917594486
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,8192,0.0274008893304401
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,7168,0.017423111531469557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,8192,0.013776889277829064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,6144,0.015712888704405892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,7168,0.026128000683254663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,7168,0.012322666744391123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,5120,0.013896889156765409
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,6144,0.02253866692384084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,6144,0.011170667078759937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,4096,0.0110880004035102
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,5120,0.00979377743270662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,5120,0.019109333554903667
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,4096,0.018765333626005385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,3584,0.0096124443742964
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,4096,0.008664888640244802
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,3072,0.008756444685988957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,3584,0.0177875558535258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,3584,0.008355555435021719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,3072,0.015850666496488783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,2560,0.008064889245563084
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,3072,0.007691555553012424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,2048,0.007662222617202335
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,2560,0.015256888336605497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,2560,0.007375111182530721
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,2048,0.013572444518407186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,1536,0.005580444302823808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,2048,0.0070186663005087115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,1536,0.013200889031092325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,1024,0.004275555411974589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,1536,0.0069004446268081665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,768,0.0038044444388813446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,1024,0.011456888582971362
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,1024,0.006268444574541516
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,512,0.0037120001183615792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,768,0.005956444475385878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,768,0.012313777373896705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,256,0.0033777778347333274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,512,0.005949333310127258
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,512,0.011780444118711682
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,128,0.0029066666546795103
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,256,0.005669333454635408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,256,0.010703999963071613
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,64,0.003010666618744532
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,5120,128,0.005660444498062134
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,5120,32,0.002893333426780171
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,128,0.0094035557574696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,64,0.009409777820110321
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,65536,0.08103289206822713
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,5120,32,0.0087013335691558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,16384,0.026361778378486633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,65536,0.07989955610699125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,65536,0.06660977999369304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,16384,0.03303111261791653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,12288,0.021564443906148274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,16384,0.02035911050107744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,10240,0.01831022236082289
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,12288,0.031375110149383545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,12288,0.01650399963061015
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,10240,0.026033777329656813
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,8192,0.016340444485346477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,10240,0.014908444550302295
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,7168,0.014648000399271647
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,8192,0.02423111100991567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,8192,0.012857777376969656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,7168,0.02440622283352746
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,6144,0.01349688900841607
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,7168,0.011788444386588203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,5120,0.012180444267061023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,6144,0.022295110755496558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,6144,0.010495999621020423
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,4096,0.011771555576059552
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,5120,0.019647111495335896
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,5120,0.009130666653315226
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,3584,0.011434666812419891
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,4096,0.008354666332403818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,3584,0.01589866644806332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,4096,0.0166284441947937
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,3072,0.01036088913679123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,3584,0.008104888929261101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,2560,0.009845332966910468
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,3072,0.007636444436179266
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,3072,0.015042665931913586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,2048,0.008735111190213097
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,2560,0.007000000112586551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,2560,0.014455111490355598
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,1536,0.007041778001520369
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,2048,0.0069448889957533936
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,2048,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,1536,0.012829333543777466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,1024,0.005997333261701796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,1536,0.006617777877383762
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,768,0.004575110971927643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,1024,0.012251555919647217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,1024,0.006275555739800136
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,768,0.011413333316644033
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,512,0.00425511101881663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,768,0.006049777898523543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,256,0.0033866665843460294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,512,0.011067555182509951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,512,0.00599911105301645
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,256,0.005612444546487596
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,128,0.00295733329322603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,256,0.01111822244193819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,64,0.0029013332807355454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,128,0.008708444734414419
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,4096,32,0.0028782222006056043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,4096,128,0.005657777604129579
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,64,0.01016266644001007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,65536,0.07342133257124159
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,4096,32,0.00868711123863856
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,16384,0.023386667172114056
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,65536,0.07651111152436998
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,65536,0.0658408866988288
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,16384,0.031493332650926374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,16384,0.020112888680564035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,12288,0.019288000133302476
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,10240,0.016840888394249808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,12288,0.029327111111746892
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,10240,0.014672888649834527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,10240,0.023385778069496155
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,8192,0.014472888575659858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,8192,0.012092444631788464
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,7168,0.013736888766288757
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,12288,0.016255999604860943
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,7168,0.022198221749729578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,8192,0.022073777185546026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,7168,0.011482666763994427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,6144,0.012858666479587555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,6144,0.020310221446885
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,6144,0.010553778045707278
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,5120,0.011301333705584208
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,5120,0.009368889033794403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,4096,0.010939555863539377
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,5120,0.017256889078352187
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,4096,0.016581333345837064
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,3584,0.011875555747085147
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,3584,0.015635554989178974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,3584,0.00814844419558843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,3072,0.010826666322019366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,3072,0.01460622251033783
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,4096,0.008026666939258575
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,2560,0.009562667045328353
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,2560,0.007288888924651676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,2048,0.008045333127180735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,2560,0.014171555638313293
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,2048,0.013473778135246701
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,1536,0.006684444430801604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,2048,0.007089777953094906
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,1536,0.012801777985360889
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,1536,0.006703111032644908
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,1024,0.005364444520738389
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,3072,0.007687999970383114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,768,0.004543111142185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,1024,0.006385777973466449
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,1024,0.011866666376590729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,512,0.004384888956944148
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,768,0.012087999946541257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,512,0.005589333259397083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,256,0.0031973332580592898
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,512,0.011755555868148804
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,256,0.01072088877360026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,128,0.0029022221763928733
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,256,0.00565244464410676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,128,0.009863110880057016
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,128,0.005647111270162795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,64,0.0028604444944196274
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3584,32,0.0028293333533737394
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3584,768,0.006045333213276333
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,32,0.010108444425794814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3584,64,0.009142222503821055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,65536,0.06367644336488512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,16384,0.021008888880411785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,65536,0.071578668223487
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,65536,0.06471288866466947
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,16384,0.019700444406933255
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,12288,0.017749332719379
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,12288,0.0266951107316547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,10240,0.015485333071814643
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,12288,0.01609422266483307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,10240,0.023545778459972803
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,10240,0.014157333307796054
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,8192,0.013774221969975365
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,8192,0.011750222080283694
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,7168,0.012595555848545499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,8192,0.020958221620983545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,16384,0.02851288848453098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,6144,0.012172444827026792
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,7168,0.021088000800874498
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,7168,0.010503110786279043
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,5120,0.010467555787828235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,6144,0.01920000049802992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,6144,0.010106667048401302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,4096,0.010637332995732626
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,5120,0.009418666362762451
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,4096,0.015526221858130561
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,3584,0.009663110805882348
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,4096,0.008403555386596257
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,3584,0.015252444479200574
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,3072,0.008557333714432186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,3072,0.01422755585776435
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,3072,0.00775999989774492
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,2560,0.007696000238259633
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,5120,0.016047999262809753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,2560,0.013849777479966482
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,2048,0.006656000183688269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,2560,0.007368000017272101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,3584,0.007997333175606197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,1536,0.007326222128338284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,2048,0.013204444613721637
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,1024,0.005234666582610872
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,2048,0.007056000332037608
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,1536,0.006984000404675801
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,768,0.00499199993080563
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,1536,0.01256622208489312
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,1024,0.006436444405052397
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,512,0.004287111262480418
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,768,0.005980444451173146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,512,0.005674666828579373
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,512,0.011767999993430244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,256,0.0036151111125946045
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,256,0.005609777652555042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,256,0.010496888723638324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,1024,0.012006222373909421
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,128,0.009415999882751042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,768,0.012493333054913415
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,64,0.0028764444092909494
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,3072,128,0.005719999886221356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,32,0.002918222298224767
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,64,0.010076444182131026
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,65536,0.05528888768619961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,3072,32,0.009460444251696268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,3072,128,0.0032035555276605818
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,16384,0.019395555059115093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,65536,0.06993777884377374
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,65536,0.06136622031529745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,16384,0.027580445011456806
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,16384,0.01940799918439653
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,12288,0.016570667425791424
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,10240,0.014152000347773233
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,12288,0.015199111567603217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,12288,0.0251920007997089
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,8192,0.012279111478063794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,10240,0.013470222552617392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,10240,0.021173333128293354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,7168,0.011328889264000786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,8192,0.021013332737816706
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,8192,0.01110400011142095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,6144,0.011500444677140979
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,7168,0.010412444670995077
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,7168,0.01860000027550591
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,5120,0.009763555394278632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,6144,0.017690666847758822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,6144,0.009734222458468543
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,4096,0.009725333087974125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,5120,0.00908622228437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,5120,0.014984000060293408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,4096,0.014160000615649752
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,3584,0.009511111511124505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,4096,0.008372444245550368
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,3072,0.007893333004580604
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,3584,0.013863999810483722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,3584,0.008055999875068665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,3072,0.013719999955760108
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,2560,0.007725333174069722
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,3072,0.007751999629868402
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,2048,0.006735111276308696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,2560,0.013585777746306526
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,2560,0.007325333025720384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,1536,0.006666666517655055
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,2048,0.012396444049146442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,2048,0.007033777733643849
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,1024,0.005014222115278244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,1536,0.00666933341158761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,1536,0.012551999754375882
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,1024,0.011920889218648275
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,1024,0.006051555689838197
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,768,0.005970666805903117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,768,0.012175111307038201
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,512,0.003912000192536248
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,512,0.01183999992079205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,512,0.0059004442559348206
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,256,0.0033866665843460294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,256,0.00944800012641483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,256,0.005685333162546158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,128,0.0032231110251612137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2560,128,0.00555377784702513
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,128,0.010618666807810465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,64,0.002862222285734283
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,768,0.0047048889100551605
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2560,32,0.002579555536309878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,65536,0.04666755596796671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,64,0.010102222363154093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2560,32,0.00942311104800966
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,65536,0.06706311305363973
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,16384,0.01665155589580536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,65536,0.05911022424697876
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,12288,0.012820444173283048
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,16384,0.02645066711637709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,16384,0.018960888187090557
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,12288,0.022797332869635686
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,10240,0.012106666962305704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,12288,0.014159111513031853
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,8192,0.010802666346232096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,10240,0.020621332857343886
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,10240,0.01291911138428582
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,8192,0.019839111301634047
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,7168,0.010139555566840703
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,8192,0.01109333336353302
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,6144,0.010406222608354356
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,7168,0.01739288866519928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,7168,0.010595555934641095
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,6144,0.01531022290388743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,5120,0.008784888519181145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,6144,0.009727111293209923
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,4096,0.009787555370065901
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,5120,0.014543111125628153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,5120,0.00943199959066179
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,4096,0.01346666696998808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,3584,0.009448889229032729
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,4096,0.008347555167145198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,3072,0.007348444726732042
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,3584,0.01348622226052814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,3584,0.008117333054542542
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,3072,0.013233777549531726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,3072,0.007682667010360294
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,2560,0.006631111105283101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,2560,0.012819555898507437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,2560,0.0073520003093613525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,2048,0.008037333687146505
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,2048,0.007071110937330458
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,2048,0.012794666820102267
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,1536,0.00675022229552269
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,1536,0.006717333363162146
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,1536,0.012191111014948951
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,1024,0.0052728888889153796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,1024,0.005978666659858491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,1024,0.011493333511882357
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,768,0.004607999904288186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,768,0.0059262220230367445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,768,0.012267555627557965
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,512,0.0038924444880750445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,512,0.01181422256761127
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,512,0.00565244464410676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,256,0.003589333345492681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,256,0.00941244430012173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,256,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,128,0.0032106666929192017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,2048,128,0.005653333332803514
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,128,0.01016266644001007
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,64,0.0028719999310043124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,2048,32,0.002875555513633622
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,64,0.01017777787314521
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,2048,32,0.008757332960764566
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,65536,0.03665688965055678
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,16384,0.014392889208263822
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,65536,0.0594115588400099
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,65536,0.057836446497175426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,12288,0.011559111376603445
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,16384,0.024890666206677754
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,12288,0.02124622298611535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,10240,0.013803555733627744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,12288,0.01388533330625958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,10240,0.018769777483410306
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,8192,0.01219911128282547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,10240,0.01257333325015174
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,8192,0.016233777006467182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,8192,0.011132444772455426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,7168,0.011384888655609555
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,7168,0.015476443701320224
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,16384,0.017869333426157635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,7168,0.010508444574144153
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,6144,0.011071110765139261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,5120,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,6144,0.009758222434255812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,6144,0.014811555544535318
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,4096,0.008639111287064022
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,5120,0.014506666196717156
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,5120,0.009022222624884712
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,3584,0.006814222368929122
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,4096,0.01292799992693795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,4096,0.00834222220712238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,3072,0.00627377794848548
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,3584,0.014179555906189812
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,3072,0.013092444174819522
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,2560,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,3072,0.007754666937722101
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,2560,0.013257777525318993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,2560,0.0073902226156658595
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,2048,0.006629333313968446
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,2048,0.007288888924651676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,2048,0.012455999851226807
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,1536,0.005590222362014983
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,1536,0.006609777609507243
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,1024,0.004519111166397731
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,3584,0.007996444072988298
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,1024,0.011800888511869641
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,768,0.0041999999019834726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,768,0.011787555283970304
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,768,0.005948444621430502
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,512,0.003656889001528422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,1536,0.012132444315486483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,512,0.011727111207114326
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,256,0.0031902222997612427
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,1024,0.005954666684071223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,256,0.01072799993885888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,128,0.002959111084540685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,256,0.005583999885453119
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,128,0.005585777676767773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,64,0.0028613333900769553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,64,0.008719111482302347
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1536,32,0.002895111011134254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1536,512,0.005619555711746216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,65536,0.026428444517983332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,32,0.008678221868144141
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,65536,0.052856889035966664
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,16384,0.015007111761305066
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1536,128,0.009697777529557547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,16384,0.021801778011851843
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,16384,0.016209777858522203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,12288,0.012471111284361945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,12288,0.013461333182122974
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,12288,0.018554665976100497
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,10240,0.011382222175598145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,10240,0.017216889394654166
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,8192,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,65536,0.05740355783038669
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,10240,0.012864000267452665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,8192,0.016269332832760282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,7168,0.00923733330435223
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,8192,0.010817777779367236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,6144,0.00865955568022198
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,7168,0.015646222564909194
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,7168,0.01015377789735794
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,6144,0.009711999860074785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,5120,0.007108444141017065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,5120,0.014489778214030795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,4096,0.007663999994595845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,5120,0.009171555439631144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,4096,0.013809777796268463
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,3584,0.007019555403126611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,4096,0.008423111504978603
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,3584,0.013820444544156393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,3072,0.006314666734801398
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,3584,0.008090666598743862
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,6144,0.014542222023010254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,2560,0.005955555372767978
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,3072,0.013077333569526672
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,3072,0.007739555504586961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,2560,0.012814222110642327
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,2560,0.007333333293596904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,2048,0.00816000004609426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,1536,0.006778666542636023
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,2048,0.0069644442862934535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,2048,0.011749332977665795
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,1024,0.004631999880075455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,1536,0.006660444455014334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,1536,0.012583110895421771
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,1024,0.01072533345884747
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,768,0.004942222187916438
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,1024,0.006278222219811545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,512,0.0038862224254343244
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,768,0.01072622256146537
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,768,0.006037333359320958
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,512,0.011759111450778114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,512,0.005592888842026393
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,256,0.00526577772365676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,256,0.010712888505723743
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,128,0.0032266666077905228
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,128,0.00978311151266098
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,64,0.0029200000895394217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,64,0.009709333380063375
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,32,0.0029057777590221832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,1024,256,0.003335111050142182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,1024,32,0.009701333112186855
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,65536,0.020974222156736586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,65536,0.05247822072770861
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,16384,0.013487999637921652
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,65536,0.057327999009026416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,1024,128,0.0052791109515561005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,12288,0.010704889065689512
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,16384,0.020608888732062448
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,16384,0.016695999436908297
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,12288,0.018258665998776753
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,10240,0.009753777749008602
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,12288,0.013548444542619916
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,8192,0.009102221992280748
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,10240,0.016750221451123554
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,10240,0.012563555604881711
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,8192,0.016221332881185744
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,7168,0.0087333329849773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,8192,0.010848888920413123
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,6144,0.008003555238246918
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,7168,0.01552088889810774
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,7168,0.010413332945770687
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,6144,0.014835556348164877
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,5120,0.006799110935793982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,6144,0.009717333647939894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,4096,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,5120,0.013807999591032663
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,5120,0.009195555415418413
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,4096,0.013847110999955071
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,3584,0.007063999772071838
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,4096,0.008225777910815345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,3072,0.0067528887755341
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,3584,0.013470222552617392
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,3584,0.008105778031879002
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,3072,0.012819555898507437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,2560,0.0063440000845326324
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,3072,0.007620444728268518
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,2048,0.0052933332820733385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,2560,0.012776888906955719
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,2560,0.007381333245171442
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,2048,0.01277511070171992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,1536,0.005714666512277391
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,2048,0.0069857777820693115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,1024,0.004594666676388847
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,1536,0.012578667038016848
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,1536,0.006688000013430913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,1024,0.010761777559916178
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,768,0.0042151109211974675
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,1024,0.006280888699822956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,512,0.003585777762863371
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,768,0.010853333605660332
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,768,0.006034666465388403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,512,0.011746666497654386
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,256,0.0032097777972618737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,512,0.005736888696750005
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,256,0.010440000229411654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,128,0.0028799999919202593
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,256,0.005576888720194499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,128,0.009717333647939894
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,64,0.00286666676402092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,768,128,0.005702222386995952
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,768,32,0.0029200000895394217
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,64,0.009413333402739631
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,65536,0.023288889063729182
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,768,32,0.009249777429633671
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,65536,0.05176711082458496
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,16384,0.010816000401973724
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,65536,0.05415199862586128
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,16384,0.02075822154680888
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,12288,0.0087013335691558
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,16384,0.01626488897535536
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,12288,0.013460444079505073
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,12288,0.01846311158604092
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,10240,0.01718222267097897
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,8192,0.008001777860853408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,10240,0.012200888660218982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,8192,0.016234666109085083
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,7168,0.008001777860853408
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,8192,0.010885333021481832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,7168,0.015503111812803479
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,7168,0.010394666757848527
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,10240,0.007978666987684038
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,6144,0.014856888188256158
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,5120,0.006702222343948152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,6144,0.009826666778988307
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,5120,0.01419911119672987
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,4096,0.0069848886794514125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,5120,0.00906844437122345
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,4096,0.013790222505728403
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,4096,0.008106666306654612
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,3584,0.0069982219073507525
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,6144,0.007307555940416124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,3072,0.006665777828958299
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,3584,0.008031999899281396
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,3584,0.013493333425786761
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,3072,0.01314755529165268
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,2560,0.0063528890411059065
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,3072,0.007701333198282454
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,2048,0.00600355573826366
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,2560,0.01331822243001726
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,2560,0.007367110914654202
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,1536,0.004663111021121343
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,2048,0.012480888929631976
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,2048,0.006948444578382704
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,1024,0.004239999999602635
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,1536,0.006709333509206772
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,1536,0.01219466659757826
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,1024,0.011804444094498953
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,768,0.003646222253640493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,1024,0.006285333385070165
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,512,0.003629333443111844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,768,0.011768889096048145
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,768,0.005717333406209946
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,512,0.01183199965291553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,512,0.005580444302823808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,256,0.005599110904667113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,256,0.011095999843544431
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,128,0.00281866660548581
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,512,128,0.005608888963858287
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,128,0.010098666780524785
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,64,0.0028071111689011254
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,32,0.00286311118139161
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,64,0.010083555347389644
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,65536,0.014879110786649915
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,512,32,0.00941244430012173
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,65536,0.04950755503442553
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,65536,0.04729333188798693
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,512,256,0.0033404444240861465
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,16384,0.008667555948098501
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,12288,0.00833422193924586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,16384,0.020404444800482858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,16384,0.016297777493794758
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,12288,0.018582221534517076
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,10240,0.00740799970097012
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,12288,0.013478221992651621
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,8192,0.007133333219422235
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,10240,0.0170151111152437
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,10240,0.01258044441541036
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,7168,0.006664888726340399
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,8192,0.011066666907734342
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,7168,0.010452444354693094
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,6144,0.006593777901596493
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,7168,0.015207111835479736
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,6144,0.014865777558750577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,5120,0.005998222364319696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,6144,0.009501333038012186
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,5120,0.014126222994592456
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,5120,0.009101333717505137
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,8192,0.0158906661801868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,4096,0.013648000028398303
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,3584,0.006339555399285422
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,4096,0.0059368887709246735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,3584,0.01349422252840466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,3072,0.006256000035338932
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,4096,0.00830666638082928
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,3584,0.007816889219813878
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,2560,0.006263111200597551
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,3072,0.01275644451379776
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,3072,0.007679111427730984
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,2048,0.005705777969625261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,2560,0.012441777520709567
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,2048,0.0069546666410234236
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,1536,0.0042168887125121225
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,1536,0.011445333560307821
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,1536,0.006317333214812809
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,1024,0.0038622220357259116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,2560,0.007313778003056844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,1024,0.010790222220950656
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,2048,0.012471111284361945
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,768,0.0035342222286595237
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,768,0.005580444302823808
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,768,0.012159110771285163
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,512,0.0032595555401510666
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,512,0.005629333357016246
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,512,0.01181244436237547
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,256,0.0029928889125585556
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,256,0.005255110975768831
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,128,0.0029031110720502008
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,1024,0.005923555543025334
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,128,0.009160888691743216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,64,0.0029075555503368378
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,256,128,0.005712888720962737
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,256,32,0.0029111111329661477
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,64,0.009036444127559662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,65536,0.013457777599493662
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,32,0.009438222481144799
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,65536,0.045978665351867676
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,65536,0.048992888795004956
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,256,256,0.010821333361996545
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,16384,0.020346666375796
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,16384,0.01620444489849938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,12288,0.007304888632562425
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,12288,0.013619555367363824
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,10240,0.006960888703664144
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,10240,0.016941333810488384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,16384,0.007983999947706858
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,10240,0.012425777812798819
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,8192,0.006581333362393909
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,8192,0.01035199976629681
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,8192,0.015823110938072205
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,7168,0.006561777657932705
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,12288,0.01822933389080895
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,7168,0.00999644481473499
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,6144,0.006355555521117316
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,6144,0.014143110977278816
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,6144,0.009744889206356471
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,5120,0.00591555568906996
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,5120,0.014176888598336114
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,4096,0.005585777676767773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,4096,0.01350933313369751
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,7168,0.015169777803950839
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,4096,0.008408000071843466
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,3584,0.006237333433495627
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,3584,0.007724444071451823
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,3584,0.013773333695199756
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,3072,0.005934222290913264
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,5120,0.008787555827034844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,3072,0.007305777735180325
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,3072,0.013167111410035027
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,2560,0.006105777704053455
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,2048,0.005655999812814925
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,2560,0.007343110938866933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,2048,0.011810666984981961
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,1536,0.0042133331298828125
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,2048,0.0069360000391801195
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,1536,0.01107911103301578
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,1536,0.006680000159475539
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,1024,0.0038684445122877755
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,1024,0.005941333456171884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,1024,0.011481777661376528
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,768,0.00360355567600992
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,768,0.005611555443869696
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,512,0.0031848889258172777
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,2560,0.013096888860066732
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,512,0.010747555229398938
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,256,0.0031724443866146933
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,512,0.005573333137565189
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,256,0.005568888866239124
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,256,0.00980177770058314
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,128,0.0031271111220121384
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,64,0.003144888828198115
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,nvfp4,1,128,128,0.005233777893914117
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,128,0.009390222529570261
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,128,32,0.0028773333049482773
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,65536,0.010375111467308467
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,32,0.009325332939624786
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,16384,0.0069777775141927935
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,768,0.012073777616024017
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,65536,0.045941332976023354
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,12288,0.006335999816656113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,128,64,0.009515555368529426
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,10240,0.005976888868543837
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,16384,0.020655999581019085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,12288,0.017649778061442904
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,8192,0.006621333460013072
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,7168,0.006682666639486949
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,10240,0.016946666770511203
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,6144,0.005965333431959152
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,8192,0.015308444698651632
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,5120,0.005933333188295364
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,6144,0.014746665954589844
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,4096,0.005335999859703912
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,5120,0.013471999930010902
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,3584,0.005631999837027655
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,4096,0.013608000344700284
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,3072,0.005953777581453323
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,3584,0.012839111189047495
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,2560,0.005614222337802251
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,3072,0.013195555243227216
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,2048,0.005985777825117111
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,2560,0.012477333347002665
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,1536,0.004275555411974589
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,2048,0.012488000094890594
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,7168,0.015266666809717814
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,1024,0.003948444293604957
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,1536,0.011189333266682096
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,768,0.0036319999231232535
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,512,0.0033155555526415506
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,1024,0.012086222569147745
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,768,0.011092444260915121
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,256,0.0032008888406885993
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,128,0.0029146667155954572
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,512,0.011737777955002256
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,64,0.0029493332323100832
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,256,0.009955555200576782
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,128,0.009713778065310584
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,64,0.010048888623714447
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,65536,0.009009777671760982
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,64,32,0.009365333451165093
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,16384,0.006246222390068903
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,65536,0.04636889033847385
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,12288,0.006011555592219035
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,16384,0.020655999581019085
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,10240,0.006100444330109491
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,12288,0.01809511085351308
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,8192,0.005554666535721884
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,64,32,0.002512000087234709
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,7168,0.005937777873542573
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,8192,0.01588000026014116
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,6144,0.005695999910434087
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,7168,0.014865777558750577
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,5120,0.005951111101441913
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,6144,0.014861333701345654
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,4096,0.005370666583379109
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,5120,0.01347644461525811
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,3584,0.005718222094906702
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,4096,0.013651555611027611
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,3072,0.005668444352017508
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,10240,0.016740444633695815
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,3584,0.012784889174832238
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,2560,0.0063235556913746735
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,2048,0.005731555736727185
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,3072,0.013232889274756113
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,1536,0.0047422221137417685
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,2560,0.012412444584899478
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,1024,0.003979555434650845
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,2048,0.012440000143316058
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,768,0.0035457776652442086
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,1536,0.012138666378127204
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,512,0.0033235556135574975
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,1024,0.010407999985747868
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,256,0.0028453332682450614
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,768,0.012132444315486483
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,128,0.0029004443850782183
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,512,0.011914666328165265
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,64,0.0029279999434947968
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,256,0.010038221875826517
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,float16,1,32,32,0.002554666664865282
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,128,0.009688888986905416
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,64,0.010064000056849586
VLLM,0.14.1.dev1+gd68209402,NVIDIA GB200,gemm,vllm_default,fp8,1,32,32,0.009000889129108852
