framework,version,device,op_name,kernel_source,mla_dtype,kv_cache_dtype,num_heads,batch_size,isl,tp_size,step,latency
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,1,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,1,0.31487999359766644
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,1,0.31197865804036456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,1,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,1,0.3495253324508667
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,1,0.340992013613383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,1,0.3454293409983317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,1,0.3433813254038493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,1,0.3213653365770976
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,1,0.3179519971211751
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,1,0.32870399951934814
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,1,0.32255999247233075
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,1,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,1,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,1,0.3362133502960205
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,1,0.3442346652348836
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,1,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,1,0.37444265683492023
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,1,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,1,0.3198293248812358
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,1,0.3218773404757182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,1,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,1,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,1,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,1,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,1,0.33177600304285687
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,1,0.35703468322753906
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,1,0.3456000089645386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,1,0.3258026639620463
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,1,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,1,0.33740798632303876
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,1,0.31573333342870075
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,1,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,1,0.3421866496404012
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,1,0.3290453354517619
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,1,0.33826132615407306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,1,0.31829333305358887
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,1,0.3242666721343994
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,1,0.3442346652348836
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,1,0.35037867228190106
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,1,0.3461120128631592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,1,0.33484800656636554
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,1,0.32290132840474445
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,1,0.31726932525634766
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,1,0.3280213276545207
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,1,0.3211946686108907
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,1,0.33075199524561566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,1,0.3322880069414775
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,1,0.3447466691335042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,1,0.34508800506591797
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,1,0.33894399801890057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,1,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,1,0.33075199524561566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,1,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,1,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,1,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,1,0.3421866496404012
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,1,0.35601067543029785
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,1,0.31641600529352826
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,1,0.37649067242940265
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,1,0.340992013613383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,1,0.33638401826222736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,1,0.32102400064468384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,1,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,1,0.5713920195897421
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,1,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,1,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,1,0.34986666838328045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,1,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,1,0.3251199920972188
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,1,0.34167468547821045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,1,0.3397973378499349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,1,0.35276798407236737
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,1,0.5601280132929484
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,1,0.8907093207041422
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,1,1.0765653451283772
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,1,0.3485013246536255
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,1,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,1,0.32443734010060626
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,1,0.33587201436360675
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,1,1.0571093559265137
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,1,0.3454293409983317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,1,0.5596160093943278
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,1,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,1,0.35447466373443604
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,1,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,1,0.33894399801890057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,1,2.082815965016683
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,3,0.3561813433965047
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,3,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,3,0.3457706769307454
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,3,0.6681599617004395
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,3,0.5178026755650839
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,3,0.355840007464091
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,3,0.35089067618052167
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,3,0.680618683497111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,3,0.3404800097147624
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,3,0.40226133664449054
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,3,0.3466240167617798
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,3,0.33604268232981366
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,3,0.343722661336263
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,3,0.3471359809239705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,3,0.3490133285522461
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,3,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,3,0.725162665049235
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,3,0.34167468547821045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,3,0.34969600041707355
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,3,0.33501867453257245
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,3,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,3,0.3201706608136495
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,3,0.33450667063395184
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,3,0.3679573138554891
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,3,0.3566933472951253
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,3,1.1060907046000164
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,3,1.6669012705485027
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,3,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,3,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,3,0.423253337542216
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,3,0.3421866496404012
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,3,0.3438933293024699
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,3,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,3,0.3438933293024699
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,3,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,3,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,3,0.33672531445821124
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,3,0.31197865804036456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,3,0.32238932450612384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,3,0.368127981821696
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,3,0.5102933247884115
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,3,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,3,0.3479893207550049
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,3,0.4648960034052531
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,3,1.4095360438028972
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,3,0.31522132953008014
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,3,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,3,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,3,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,3,0.3457706769307454
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,3,0.34167468547821045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,3,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,3,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,3,0.3211946686108907
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,3,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,3,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,3,0.33655468622843426
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,3,0.37614933649698895
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,3,0.851967970530192
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,3,1.2917760213216145
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,3,0.8215893109639486
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,3,0.37887998421986896
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,3,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,3,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,3,0.3590826590855916
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,3,0.3394560019175212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,3,0.47086934248606366
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,3,0.6272000074386597
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,3,0.8758613268534342
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,3,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,3,0.3694933255513509
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,3,0.30668799082438153
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,3,0.4804266691207886
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,3,0.5374293327331543
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,3,0.8773972988128662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,3,0.3466240167617798
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,3,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,3,1.7017173767089844
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,3,0.3114666740099589
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,3,0.3259733319282532
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,3,0.49186134338378906
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,3,0.3643733263015747
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,3,0.8833706378936768
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,3,0.35788798332214355
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,3,1.6837973594665527
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,7,0.32102400064468384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,3,3.3522345225016275
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,7,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,7,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,7,0.3510613441467285
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,7,0.6582613388697306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,3,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,7,0.32921600341796875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,7,1.01256529490153
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,7,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,7,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,7,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,7,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,7,0.3380906581878662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,3,0.36300798257191974
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,7,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,7,1.0142719745635986
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,7,1.3213013013203938
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,7,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,7,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,7,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,7,0.33536001046498615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,7,0.32409600416819256
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,7,1.6928426424662273
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,7,1.8570240338643391
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,7,0.3546453317006429
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,7,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,7,0.32255999247233075
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,7,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,7,0.3444053332010905
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,7,0.673962672551473
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,7,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,7,1.398954709370931
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,7,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,7,0.8091306686401367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,7,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,7,0.5321386655171713
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,7,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,7,0.8756906986236572
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,7,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,7,0.340992013613383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,7,0.3609600067138672
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,7,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,7,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,7,0.33672531445821124
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,7,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,7,0.3404800097147624
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,7,0.5341866811116537
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,7,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,7,1.2837546666463215
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,7,0.3449173370997111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,7,0.33826132615407306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,7,0.8488960266113281
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,7,0.34679468472798664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,7,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,7,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,7,0.3256319959958394
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,7,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,7,0.3397973378499349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,7,0.3546453317006429
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,7,0.5573973258336385
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,7,0.6983679930369059
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,7,0.34406399726867676
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,7,0.5492053429285685
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,7,0.33484800656636554
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,7,0.3107840021451314
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,7,0.3170986572901408
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,7,1.072981357574463
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,7,0.5591040054957072
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,7,0.5777066548665365
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,7,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,7,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,7,1.0362880229949951
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,7,0.46114134788513184
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,7,0.3193173408508301
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,7,0.35072000821431476
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,7,1.38427734375
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,7,0.562175989151001
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,7,0.3519146839777629
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,7,2.05841064453125
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,7,1.0367999871571858
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,7,0.3597653309504191
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,7,0.3421866496404012
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,7,0.41625599066416424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,7,1.0395306746164958
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,7,0.35942399501800537
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,7,0.5640533367792765
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,7,0.35498666763305664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,7,0.5522773265838623
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,7,4.02619743347168
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,7,2.0157440503438315
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,15,0.3131733338038127
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,15,0.3213653365770976
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,15,0.5631999969482422
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,15,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,15,0.34355199337005615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,15,0.3553280035654704
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,15,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,15,0.34167468547821045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,15,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,15,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,15,0.337066650390625
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,15,0.33058132727940875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,15,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,15,0.6166186730066935
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,15,0.3397973378499349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,15,0.3298986752827962
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,15,0.3394560019175212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,15,0.8582826455434164
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,15,0.33638401826222736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,15,0.33723731835683185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,15,0.3485013246536255
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,15,0.7557120323181152
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,15,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,15,0.3394560019175212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,15,0.3659093379974365
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,15,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,15,0.33740798632303876
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,15,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,15,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,15,2.1724160512288413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,15,0.6910293102264404
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,15,0.33655468622843426
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,15,0.7239680290222168
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,15,1.0815146764119465
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,15,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,15,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,15,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,15,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,15,1.9462827046712239
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,15,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,15,0.3237546682357788
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,15,0.33501867453257245
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,15,0.33604268232981366
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,15,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,15,0.6396586497624716
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,15,0.764415979385376
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,15,0.3314346671104431
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,15,1.005226691563924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,15,0.33348266283671063
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,15,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,15,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,15,0.33723731835683185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,15,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,15,0.33723731835683185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,15,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,15,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,15,0.5599573453267416
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,15,0.3394560019175212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,15,0.7731200059254965
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,15,0.33672531445821124
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,15,0.33655468622843426
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,15,0.33774932225545246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,15,0.35259731610616046
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,15,0.31624533732732135
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,15,0.6067200104395548
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,15,1.0432853698730469
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,15,0.5597866773605347
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,15,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,15,0.5160959959030151
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,15,0.52292267481486
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,15,0.32341333230336505
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,15,0.31436800956726074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,15,0.4459520181020101
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,15,1.0574506918589275
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,15,0.5758293469746908
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,15,0.3454293409983317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,15,0.5133653481801351
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,15,2.0602879524230957
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,15,0.3575466473897298
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,15,0.3198293248812358
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,15,1.0530133247375488
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,15,0.575488011042277
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,15,0.32972800731658936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,31,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,15,0.4256426493326823
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,31,0.36232535044352215
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,31,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,15,4.081151962280273
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,31,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,31,0.3324586749076843
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,31,0.3515733480453491
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,31,0.3300693432490031
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,31,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,31,0.9944746494293213
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,31,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,31,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,31,0.31641600529352826
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,31,1.0197333494822185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,31,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,31,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,31,0.3810986677805583
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,31,0.33638401826222736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,31,0.33484800656636554
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,31,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,31,0.3380906581878662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,31,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,31,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,31,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,31,1.020416021347046
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,31,0.33450667063395184
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,31,0.33501867453257245
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,31,0.34508800506591797
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,31,0.33774932225545246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,31,0.33450667063395184
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,31,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,15,0.36164267857869464
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,31,0.8198826313018799
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,31,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,31,0.33484800656636554
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,31,0.3333119948705037
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,15,2.6127360661824546
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,31,0.33484800656636554
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,15,0.4220586617787679
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,31,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,31,0.3476479848225911
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,31,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,31,0.31624533732732135
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,31,0.31539199749628705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,31,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,31,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,31,1.668949286142985
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,31,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,31,1.662293275197347
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,31,0.33348266283671063
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,31,0.32290132840474445
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,31,0.32392533620198566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,31,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,31,0.6976853211720785
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,31,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,31,0.3566933472951253
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,31,1.8445653915405273
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,31,0.7794346809387207
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,31,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,31,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,31,0.560640017191569
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,31,0.856234629948934
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,31,1.5259307225545247
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,31,0.343722661336263
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,31,0.34628268082936603
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,31,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,31,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,31,0.3118079900741577
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,31,0.3362133502960205
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,31,1.0345813433329265
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,31,1.0516479810078938
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,31,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,31,0.7932586669921875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,31,0.5782186587651571
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,31,0.3141973416010539
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,31,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,31,0.5732693274815878
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,31,0.3561813433965047
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,31,1.5329279899597168
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,31,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,31,1.424554665883382
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,31,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,31,2.079744021097819
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,31,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,31,0.59170134862264
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,31,0.36232535044352215
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,31,1.059157371520996
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,31,0.3563520113627116
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,63,0.3145386576652527
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,31,0.47018667062123615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,31,2.058922608693441
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,31,4.089173316955566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,63,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,63,0.33774932225545246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,63,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,63,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,63,0.36061867078145343
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,63,0.3375786542892456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,63,1.0714453061421711
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,63,0.36829864978790283
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,63,0.9321813583374023
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,63,0.4333226680755615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,63,0.31948800881703693
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,63,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,63,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,63,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,63,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,63,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,63,0.6787412961324056
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,63,0.5474986632664999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,63,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,63,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,63,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,63,1.0135893026987712
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,63,0.33928533395131427
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,63,1.3725013732910156
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,63,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,63,0.33604268232981366
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,63,0.7623679637908936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,63,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,63,0.3457706769307454
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,63,0.3319466710090637
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,63,0.8907093207041422
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,63,0.951637347539266
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,63,0.6359039942423502
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,63,1.0098346869150798
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,63,2.3077546755472818
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,63,0.3461120128631592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,63,0.3529386520385742
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,63,0.33501867453257245
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,63,0.33348266283671063
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,63,1.4315519332885742
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,63,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,63,0.3447466691335042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,63,0.5309439897537231
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,63,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,63,0.4580693244934082
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,63,0.3456000089645386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,63,0.44697598616282147
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,31,0.9053866863250732
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,63,0.3380906581878662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,63,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,63,0.35601067543029785
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,63,1.0106879870096843
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,63,1.025877316792806
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,63,0.5133653481801351
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,63,0.5193386475245158
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,63,0.3375786542892456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,63,1.6815786361694336
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,63,0.3597653309504191
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,63,0.5215573310852051
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,63,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,63,0.5613226493199667
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,63,0.33075199524561566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,63,0.42393600940704346
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,63,0.3380906581878662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,63,0.33826132615407306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,63,0.5765120188395182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,63,0.36130134264628094
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,63,0.34406399726867676
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,63,1.0804906686147053
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,63,0.337066650390625
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,63,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,63,0.31692800919214886
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,63,0.3479893207550049
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,63,1.9092480341593425
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,63,0.5376000006993612
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,63,0.38519465923309326
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,63,1.0898773670196533
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,63,2.061482588450114
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,63,0.3242666721343994
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,63,0.353279987970988
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,63,1.0542079607645671
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,63,0.43008001645406085
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,63,2.04697593053182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,63,0.3536213239034017
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,63,0.36164267857869464
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,63,4.05401611328125
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,63,0.567466656366984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,63,0.5550080140431722
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,127,0.3314346671104431
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,127,0.4761600097020467
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,127,0.3551573355992635
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,127,0.37137067317962646
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,127,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,127,0.3449173370997111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,127,0.9333759943644205
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,127,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,127,0.47308798631032306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,127,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,127,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,127,0.3399680058161418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,127,1.2784639994303386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,127,0.7427413463592529
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,127,0.35089067618052167
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,127,0.3449173370997111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,127,0.3269973397254944
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,127,1.0199039777119954
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,127,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,127,0.35225598017374676
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,127,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,127,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,127,0.4457813501358032
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,127,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,127,0.35549867153167725
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,127,0.33553067843119305
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,127,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,127,2.309973398844401
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,127,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,127,0.31590400139490765
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,127,0.34167468547821045
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,127,0.3438933293024699
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,127,0.33075199524561566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,127,0.6681599617004395
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,127,0.3237546682357788
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,127,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,127,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,127,0.3278506596883138
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,127,0.3341653347015381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,127,0.3386026620864868
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,127,0.429909348487854
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,127,0.3394560019175212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,127,0.3653973340988159
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,127,0.3283626635869344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,127,0.3189760049184163
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,127,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,127,0.31829333305358887
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,127,0.3362133502960205
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,127,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,127,0.5739519993464152
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,127,0.3510613441467285
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,127,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,127,0.3346773386001587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,127,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,127,0.33894399801890057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,127,0.33672531445821124
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,127,0.3176106611887614
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,127,1.0828800201416016
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,127,0.6079146862030029
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,127,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,127,0.3186346689860026
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,127,0.32102400064468384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,127,0.32477867603302
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,127,0.32102400064468384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,127,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,127,0.3237546682357788
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,127,0.3259733319282532
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,127,0.32494932413101196
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,127,0.32358400026957196
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,127,0.5901653369267782
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,127,1.1043840249379475
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,127,0.31726932525634766
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,127,2.156031926472982
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,127,0.3423573176066081
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,127,0.36693334579467773
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,127,0.34696535269419354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,127,0.32716800769170123
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,127,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,127,0.3983360131581624
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,127,4.254207928975423
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,127,1.1006293296813965
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,127,0.5800960063934326
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,127,0.3763200044631958
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,127,2.133845329284668
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,255,0.31675734122594196
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,255,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,255,0.33689598242441815
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,255,0.3300693432490031
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,127,0.45738665262858075
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,255,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,255,0.3474773168563843
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,255,0.7637333075205485
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,255,0.3155626654624939
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,255,0.32972800731658936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,255,0.33638401826222736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,255,0.3331413269042969
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,255,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,255,0.3466240167617798
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,255,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,127,0.3671040137608846
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,255,1.6791893641153972
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,255,1.0088106791178386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,255,0.3179519971211751
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,255,0.3227306604385376
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,255,0.34201598167419434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,255,0.3447466691335042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,127,0.34679468472798664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,255,0.7007573445638021
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,255,0.3336533308029175
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,255,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,255,0.33177600304285687
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,255,0.3375786542892456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,255,0.33894399801890057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,255,1.3474133809407551
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,255,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,255,1.5788373947143555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,255,1.3888853391011555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,255,0.3456000089645386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,255,0.340992013613383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,255,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,255,0.340992013613383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,255,0.8936106363932291
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,255,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,255,0.34679468472798664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,255,0.3186346689860026
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,255,0.34833065668741864
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,255,0.32631466786066693
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,255,1.2115626335144043
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,255,0.33655468622843426
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,255,0.8669866720835367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,255,1.1424427032470703
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,255,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,255,0.34935466448465985
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,255,0.33536001046498615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,127,0.7133866945902506
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,255,0.3179519971211751
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,255,0.3973120053609212
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,255,0.3536213239034017
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,255,0.8639146486918131
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,255,1.5010132789611816
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,255,0.35037867228190106
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,255,2.303317387898763
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,255,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,255,0.3391146659851074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,255,0.39236267407735187
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,255,0.7125333150227865
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,255,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,255,1.5064746538798015
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,255,0.33774932225545246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,255,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,255,0.32238932450612384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,255,0.3304106593132019
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,255,0.7444480260213217
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,255,0.3942399819691976
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,255,0.32255999247233075
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,255,1.371989409128825
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,255,0.3457706769307454
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,255,1.0154666900634766
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,255,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,255,0.33740798632303876
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,255,0.7296000321706136
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,255,2.7185494105021157
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,255,0.4020906686782837
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,255,0.3587413231531779
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,255,1.3844480514526367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,255,0.36300798257191974
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,255,0.4792319933573405
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,255,0.36232535044352215
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,255,0.5889706611633301
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,255,0.7210666338602701
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,255,1.0868053436279297
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,511,0.517632007598877
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,255,1.4011732737223308
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,255,2.69926389058431
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,255,1.382570743560791
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,255,0.8809813658396403
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,511,0.3397973378499349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,511,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,511,0.3380906581878662
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,511,0.3461120128631592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,511,0.34815998872121173
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,255,5.3758290608723955
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,511,0.35310932000478107
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,511,0.874837319056193
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,511,0.5643946727116903
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,511,0.3452586730321248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,511,0.6190079847971598
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,511,0.3473066488901774
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,511,0.35259731610616046
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,511,0.3471359809239705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,511,0.33177600304285687
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,511,0.3444053332010905
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,511,0.3433813254038493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,511,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,511,0.9308160146077474
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,511,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,511,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,511,0.34406399726867676
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,511,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,511,1.7704960505167644
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,511,0.3456000089645386
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,511,0.3375786542892456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,511,0.3211946686108907
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,511,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,511,0.3433813254038493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,511,0.5812906821568807
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,511,1.3704533576965332
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,511,0.3357013463973999
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,511,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,511,0.3824640115102132
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,511,0.7043413321177164
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,511,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,511,0.8872959613800049
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,511,0.33399466673533124
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,511,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,511,0.3461120128631592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,511,0.38707200686136883
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,511,1.5616000493367512
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,511,0.37307735284169513
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,511,0.35276798407236737
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,511,0.3853653271993001
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,511,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,511,0.32921600341796875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,511,0.31965865691502887
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,511,0.3505493402481079
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,511,0.33791999022165936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,511,0.8608427047729492
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,511,0.5306026538213094
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,511,0.3362133502960205
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,511,0.3397973378499349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,511,0.8483839829762777
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,511,0.343722661336263
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,511,0.3421866496404012
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,511,0.7637333075205485
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,511,1.1639466285705566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,511,0.9927679697672526
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,511,0.6744746367136637
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,511,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,511,0.6449493169784546
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,511,0.5314559936523438
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,511,1.9700053532918294
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,511,1.3564586639404297
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,511,0.3519146839777629
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,511,1.0961919625600178
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,511,1.0275839964548747
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,511,1.0444800059000652
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,511,0.8936106363932291
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,511,0.4766720136006673
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,511,0.8731306393941244
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,511,0.46916266282399494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,511,1.9604479471842449
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,511,1.023146629333496
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,511,0.6898346741994222
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,511,0.43195732434590656
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,511,0.5468159914016724
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,511,3.865088144938151
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,511,1.9548160235087078
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,511,1.3276159763336182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,511,1.6465919812520344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,511,0.9844053586324056
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,511,0.8523093064626058
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,511,1.0144426822662354
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,511,3.909290631612142
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,511,7.669418970743815
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,1023,0.33211733897527057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,1023,0.32614399989446
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,1023,0.3184640010197957
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,1023,0.32392533620198566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,1023,0.3628373146057129
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,1023,0.34679468472798664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,1023,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,1023,0.3601066668828328
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,1023,0.34781865278879803
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,1023,0.32631466786066693
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,1023,0.3415040175120036
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,1023,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,1023,1.6127999623616536
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,1023,0.343722661336263
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,1023,0.3218773404757182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,1023,0.3217066725095113
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,1023,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,1023,0.3413333495457967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,1023,0.35037867228190106
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,1023,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,1023,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,1023,0.3954346577326457
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,1023,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,1023,0.3264853358268738
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,1023,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,1023,0.4317866563796997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,1023,0.33740798632303876
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,1023,0.34252798557281494
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,1023,0.3232426643371582
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,1023,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,1023,0.3298986752827962
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,1023,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,1023,0.3426986535390218
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,1023,0.3408213456471761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,1023,0.3473066488901774
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,1023,0.45482667287190753
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,1023,0.33587201436360675
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,1023,0.3351893424987793
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,1023,0.32733867565790814
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,1023,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,1023,0.3242666721343994
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,1023,0.33587201436360675
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,1023,0.7963306903839111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,1023,0.34833065668741864
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,1023,0.45073068141937256
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,1023,0.33689598242441815
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,1023,0.353279987970988
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,1023,0.33843199412027997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,1023,0.33450667063395184
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,1023,1.6023893356323242
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,1023,0.8372906843821207
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,1023,0.46984533468882245
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,1023,0.3403093417485555
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,1023,0.3375786542892456
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,1023,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,1023,0.33126399914423627
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,1023,0.3264853358268738
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,1023,3.1144959131876626
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,1023,1.5964159965515137
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,1023,0.8587946891784668
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,1023,0.4652373393376668
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,1023,0.4154026508331299
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,1023,0.39560532569885254
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,1023,0.38707200686136883
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,1023,0.8724479675292969
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,1023,0.32443734010060626
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,1023,0.3298986752827962
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,1023,0.32716800769170123
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,1023,6.14024543762207
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,1023,0.3278506596883138
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,1023,3.1020374298095703
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,1023,0.33023999134699505
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,1023,0.8381439844767252
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,1023,1.596245288848877
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,1023,0.7394986947377523
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,1023,0.711680014928182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,1023,0.6778879960378011
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,1023,1.6546133359273274
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,1023,0.36369065443674725
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,1023,3.103573481241862
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,1023,6.109525044759114
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,2047,0.32341333230336505
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,1023,0.5290666818618774
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,2047,0.37461332480112713
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,2047,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,1023,12.212053934733072
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,2047,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,1023,1.5950506528218586
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,2047,0.35276798407236737
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,1023,1.3893973032633464
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,1023,1.3054292996724446
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,1023,1.2759040196736653
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,2047,0.34457600116729736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,2047,0.34884266058603924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,2047,0.605183998743693
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,2047,0.34628268082936603
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,2047,0.35072000821431476
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,2047,1.017514705657959
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,2047,0.34116268157958984
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,2047,0.3490133285522461
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,2047,0.34781865278879803
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,2047,0.35703468322753906
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,2047,0.478549321492513
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,2047,2.009258588155111
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,2047,0.35089067618052167
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,2047,0.8202239672342936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,2047,0.339626669883728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,2047,0.3471359809239705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,2047,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,2047,0.38075733184814453
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,2047,1.7300480206807454
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,2047,0.36300798257191974
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,2047,0.3901439905166626
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,2047,0.3447466691335042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,2047,0.3217066725095113
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,2047,0.34679468472798664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,2047,0.6012586752573649
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,2047,0.3461120128631592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,2047,0.35140268007914227
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,2047,0.3505493402481079
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,2047,0.35140268007914227
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,2047,0.41915734608968097
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,2047,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,2047,0.3466240167617798
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,2047,0.3418453137079875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,2047,0.34065067768096924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,1023,3.161087989807129
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,2047,0.4106239875157674
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,2047,0.45533867677052814
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,2047,0.3781973520914714
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,2047,0.3592533270517985
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,2047,0.7430826822916666
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,2047,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,2047,0.7007573445638021
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,2047,0.3232426643371582
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,2047,0.3725653489430745
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,2047,0.42018131415049237
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,2047,0.7449599901835123
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,2047,1.3655039469401042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,2047,0.35072000821431476
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,2047,0.9291093349456787
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,2047,0.35396265983581543
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,2047,0.6635520060857137
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,2047,0.3454293409983317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,2047,0.7707306543986002
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,2047,0.3850239912668864
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,2047,1.4177279472351074
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,2047,0.3723946809768677
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,2047,2.7477334340413413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,2047,0.7611733277638754
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,2047,0.37614933649698895
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,2047,0.9931093056996664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,2047,0.7033173243204752
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,2047,1.3687465985616047
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,2047,0.6761813163757324
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,2047,1.4312106768290203
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,2047,0.6676479975382487
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,2047,2.7511466344197593
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,2047,5.3942610422770185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,2047,1.6523946126302083
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,2047,1.276586691538493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,2047,1.4204586346944172
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,2047,1.2361386617024739
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,2047,2.7388585408528647
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,2047,5.374122619628906
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,2047,1.2122453053792317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,2047,10.708651224772135
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,2047,3.207680066426595
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,2047,2.37772798538208
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,2047,10.672640482584635
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,4095,0.9169920285542806
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,4095,0.3524266481399536
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,2047,21.28384017944336
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,4095,0.3519146839777629
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,4095,0.35788798332214355
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,4095,1.5455573399861653
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,4095,0.9263786474863688
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,4095,0.9120426972707113
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,4095,0.9173333644866943
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,4095,0.3566933472951253
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,4095,0.353279987970988
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,4095,0.34594134489695233
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,4095,1.1535360018412273
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,4095,0.3473066488901774
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,4095,1.405610720316569
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,4095,0.34013867378234863
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,4095,2.2155946095784507
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,4095,1.70905605951945
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,4095,0.3490133285522461
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,4095,0.3563520113627116
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,2047,2.4528212547302246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,4095,0.3433813254038493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,4095,0.3433813254038493
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,4095,0.34508800506591797
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,4095,0.3285333315531413
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,4095,0.3809279998143514
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,4095,0.45397333304087323
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,4095,0.362496018409729
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,4095,1.0219519933064778
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,4095,0.34355199337005615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,4095,0.34628268082936603
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,4095,0.5855573415756226
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,4095,0.49544533093770343
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,4095,0.5814613501230875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,2047,5.3942610422770185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,4095,0.7051946322123209
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,4095,0.3428693215052287
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,4095,0.3677866856257121
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,4095,0.44544001420338947
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,4095,0.3314346671104431
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,4095,0.3432106574376424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,4095,0.34355199337005615
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,4095,0.6879573663075765
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,4095,1.3166933059692383
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,4095,0.6999039649963379
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,2047,2.3531519571940103
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,4095,1.1033600171407063
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,4095,0.3998719851175944
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,4095,0.3495253324508667
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,2047,2.87607479095459
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,4095,0.36181334654490155
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,4095,0.3601066668828328
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,4095,0.6606506506601969
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,2047,6.191616058349609
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,4095,1.3231786886850994
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,4095,0.41625599066416424
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,4095,2.4683519999186196
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,4095,0.4623359839121501
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,4095,0.4235946734746297
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,4095,0.7164586385091146
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,4095,0.4123306671778361
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,4095,1.285973310470581
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,4095,0.733354647954305
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,4095,0.6691839694976807
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,4095,1.3784747123718262
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,4095,0.6568959951400757
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,4095,2.577749411265055
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,4095,5.061461448669434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,4095,0.6541653474171957
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,4095,1.8962772687276204
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,4095,2.5574399630228677
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,4095,1.3824000358581543
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,4095,1.2579840024312336
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,4095,1.2366507053375244
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,4095,5.060266812642415
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,4095,1.2298239866892497
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,4095,9.980586369832357
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,4095,3.1914666493733725
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,4095,9.968810399373373
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,4095,2.2958079973856607
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,4095,2.3145813941955566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,4095,2.357759952545166
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,4095,5.048661231994629
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,4095,2.596351941426595
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,4095,6.31108283996582
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,4095,19.91424051920573
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,4095,5.081941286722819
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,4095,19.789994557698567
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,8191,0.3597653309504191
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,4095,9.961983998616537
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,8191,0.3985066811243693
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,8191,0.33877333005269367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,8191,0.36266668637593585
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,4095,4.603562672932942
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,8191,0.3614720106124878
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,8191,0.36130134264628094
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,4095,4.525568008422852
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,8191,0.3561813433965047
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,8191,0.9654613335927328
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,8191,0.3725653489430745
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,8191,0.3447466691335042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,8191,0.3686399857203166
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,8191,0.36420265833536786
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,8191,0.45875199635823566
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,8191,0.35225598017374676
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,8191,0.3595946629842122
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,4095,4.501845359802246
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,8191,0.6667946974436442
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,8191,0.3660800059636434
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,8191,0.5543253421783447
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,8191,0.37768534819285077
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,8191,0.368127981821696
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,8191,0.35839998722076416
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,4095,39.52520497639974
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,8191,0.36061867078145343
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,4095,12.311040242513021
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,8191,0.6562133232752482
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,8191,1.39793062210083
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,8191,0.5468159914016724
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,8191,0.7913813591003418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,8191,1.0412373542785645
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,8191,0.3662506739298503
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,8191,1.0248533089955647
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,8191,0.36403199036916095
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,8191,1.0980693499247234
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,8191,0.36505599816640216
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,8191,0.8290987014770508
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,8191,0.7616853713989258
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,8191,1.0897066593170166
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,8191,0.5527893304824829
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,8191,0.3712000052134196
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,8191,1.279146671295166
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,8191,0.3573760191599528
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,8191,1.1187199751536052
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,8191,1.3276159763336182
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,8191,1.0048853556315105
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,8191,2.4712533950805664
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,8191,0.5196799834569296
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,8191,1.2798293431599934
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,8191,0.5125120083491007
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,8191,1.2487680117289226
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,8191,0.5210453271865845
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,8191,0.7529813448588053
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,8191,0.9096533457438151
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,8191,2.496511936187744
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,8191,0.7635626792907715
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,8191,0.7565653324127197
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,8191,4.679168065388997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,8191,1.3189120292663574
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,8191,2.431829293568929
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,8191,1.238869349161784
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,8191,1.3697706858317058
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,8191,2.5780906677246094
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,8191,1.2296533584594727
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,8191,1.2293120225270588
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,8191,4.874581336975098
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,8191,9.642154693603516
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,8191,3.7067092259724936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,8191,4.869802792867024
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,8191,2.6100053787231445
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,8191,9.629525502522787
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,8191,2.390186627705892
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,8191,2.364415963490804
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,8191,2.4009386698404946
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,8191,19.214677174886067
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,8191,6.322175979614258
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,8191,4.519935925801595
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,8191,9.645055770874023
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,8191,4.920490582784017
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,8191,4.483071962992351
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,8191,4.455423990885417
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,8191,19.135147094726562
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,8191,38.28138732910156
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,8191,12.499797821044922
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,16383,0.5459626515706381
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,16383,0.5505706469217936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,16383,0.5512533187866211
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,16383,0.5654186805089315
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,16383,0.5343573490778605
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,16383,0.532309333483378
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,16383,0.5292373498280843
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,16383,0.4720640182495117
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,16383,0.4657493432362874
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,16383,0.4671146472295125
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,16383,1.0033493041992188
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,16383,2.0061866442362466
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,16383,0.5428906679153442
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,16383,0.5394773483276367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,16383,0.5340160131454468
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,16383,0.4773546854654948
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,16383,0.47377065817515057
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,16383,0.47172268231709796
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,16383,1.473365306854248
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,16383,2.062335968017578
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,16383,1.0045440196990967
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,16383,0.5725866556167603
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,16383,0.5464746554692587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,16383,0.48349865277608234
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,16383,0.4797439972559611
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,16383,0.9937919775644938
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,16383,2.1336746215820312
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,16383,2.411349296569824
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,16383,1.0103466510772705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,16383,1.4726826349894206
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,16383,0.505514661471049
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,16383,0.5577386617660522
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,16383,0.502784013748169
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,16383,0.5022720098495483
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,16383,0.48213334878285724
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,16383,2.1792426109313965
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,16383,2.437290668487549
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,16383,4.773887952168782
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,16383,1.4878719647725422
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,16383,1.0412373542785645
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,16383,0.9659732977549235
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,16383,0.9487360318501791
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,16383,0.9545386632283529
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,16383,2.422442595163981
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,16383,0.43537068367004395
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,16383,0.5423786640167236
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,16383,2.491221268971761
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,16383,4.832085291544597
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,16383,1.5503360430399578
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,16383,1.4400854110717773
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,16383,1.4312106768290203
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,16383,1.425920009613037
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,16383,9.093973159790039
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,16383,1.0871466795603435
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,16383,4.7457278569539385
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,16383,1.244330644607544
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,16383,2.57041072845459
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,16383,2.453162670135498
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,16383,2.3558826446533203
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,16383,4.996949195861816
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,16383,2.569045384724935
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,16383,9.47268295288086
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,16383,18.957482655843098
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,16383,18.863445281982422
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,16383,5.0430294672648115
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,16383,7.312725067138672
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,16383,9.479509353637695
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,16383,4.62609068552653
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,16383,4.613290786743164
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,16383,37.58728535970052
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,16383,4.663296063741048
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,16383,12.49826176961263
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,32767,0.9925973415374756
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,32767,0.9840640226999918
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,32767,0.9770666758219401
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,32767,0.9845759868621826
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,32767,0.8669866720835367
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,32767,0.8605013688405355
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,32767,0.8610133330027262
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,32767,0.9937919775644938
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,32767,3.903317451477051
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,32767,0.9826986789703369
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,32767,0.986624002456665
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,32767,0.9828693072001139
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,32767,0.868010679880778
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,32767,0.8657920360565186
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,32767,0.8628906408945719
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,32767,1.9017386436462402
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,32767,3.952639897664388
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,32767,0.9917439619700114
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,32767,0.9954986572265625
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,32767,0.9953280289967855
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,32767,0.8719360033671061
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,32767,0.8765439987182617
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,32767,0.8715946674346924
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,32767,4.11733341217041
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,32767,2.8214613596598306
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,32767,1.9119787216186523
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,32767,0.8890026410420736
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,32767,0.8866133689880371
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,32767,0.8852480252583822
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,32767,4.183210690816243
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,32767,4.6764373779296875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,32767,2.851327896118164
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,32767,1.9273386001586914
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,32767,1.0287786324818928
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,32767,0.9296212991078695
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,32767,0.9282560348510742
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,32767,0.9265493551890055
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,32767,4.267690658569336
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,32767,4.752213478088379
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,32767,1.8524160385131836
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,32767,9.334271748860678
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,32767,1.8341546058654785
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,32767,4.712106704711914
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,32767,9.60153579711914
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,32767,18.39513651529948
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,32767,4.863146781921387
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,32767,3.0300159454345703
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,32767,2.7905705769856772
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,32767,2.773162523905436
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,32767,2.758485476175944
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,32767,9.481215794881185
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,32767,18.899796803792317
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,32767,37.17597961425781
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,32767,9.803946812947592
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,32767,5.001215934753418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,32767,4.664661407470703
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,32767,4.6074879964192705
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,32767,4.619434674580892
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,32767,14.504277547200521
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,65535,1.8863786061604817
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,65535,1.898837407430013
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,65535,1.6643412907918294
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,65535,1.6571733156840007
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,65535,1.658880074818929
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,65535,7.753557205200195
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,65535,1.9048105875651042
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,65535,1.894741376241048
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,65535,1.896447976430257
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,65535,1.9060053825378418
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,65535,1.6727040608723958
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,65535,1.6667307217915852
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,65535,1.664511998494466
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,65535,7.842133204142253
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,65535,3.7188266118367515
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,65535,1.9128319422403972
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,65535,1.9140267372131348
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,65535,1.9208532969156902
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,65535,1.68669859568278
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,65535,1.6802132924397786
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,65535,1.6807252566019695
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,65535,8.16162109375
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,65535,5.5574188232421875
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,65535,3.743232091267904
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,65535,1.9858773549397786
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,65535,1.9568640391031902
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,65535,1.714687983194987
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,65535,1.70632537206014
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,65535,1.7080319722493489
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,65535,8.343893051147461
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,65535,9.275221506754557
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,65535,5.613738377888997
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,65535,3.7867520650227866
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,65535,1.9817813237508137
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,65535,1.795413335164388
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,65535,1.7906346321105957
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,65535,1.7914880116780598
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,65535,8.466090520222982
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,65535,18.664960225423176
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,65535,9.393152236938477
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,65535,5.69053840637207
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,65535,3.9161173502604165
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,65535,3.6703573862711587
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,65535,3.639808019002279
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,65535,3.6259838740030923
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,65535,9.411754608154297
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,65535,36.988756815592446
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,65535,19.321856180826824
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,65535,9.68004290262858
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,65535,5.932885487874349
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,65535,5.536597569783528
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,65535,18.89843241373698
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,131071,3.6906665166219077
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,131071,3.2549546559651694
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,131071,3.2378880182902017
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,131071,3.670186678568522
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,131071,15.396692911783854
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,131071,3.6817919413248696
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,131071,3.6899840037027993
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,131071,3.7053438822428384
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,131071,3.352405230204264
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,131071,3.2394240697224936
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,131071,3.2544425328572593
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,131071,3.803647994995117
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,131071,7.314261118570964
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,131071,3.738453229268392
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,131071,3.74835205078125
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,131071,15.628799438476562
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,131071,3.2954025268554688
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,131071,3.2824319203694663
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,131071,3.271338780721029
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,131071,10.957995096842447
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,131071,16.279552459716797
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,131071,3.8879572550455728
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,131071,7.348735809326172
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,131071,3.8142293294270835
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,131071,4.014933268229167
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,131071,3.330730756123861
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,131071,3.3134934107462564
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,131071,16.641194661458332
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,131071,11.083946228027344
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,131071,18.384384155273438
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,131071,3.8667945861816406
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,131071,7.469738642374675
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,131071,3.4884265263875327
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,131071,3.5037867228190103
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,131071,3.4882561365763345
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,131071,3.727018674214681
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,131071,3.23635196685791
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,131071,3.6858879725138345
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,131071,16.669867197672527
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,131071,11.220480600992838
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,131071,7.739562352498372
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,131071,18.772139231363933
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,131071,7.236437479654948
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,131071,7.157930374145508
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,131071,7.096490859985352
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,131071,37.44494883219401
VLLM,0.12.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,131071,18.68236796061198
