framework,version,device,op_name,kernel_source,mla_dtype,kv_cache_dtype,num_heads,batch_size,isl,tp_size,step,latency
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,1,0.31142399708429974
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,1,0.30188266436258954
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,1,0.3020533323287964
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,1,0.3012266755104065
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,1,0.2977919975916545
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,1,0.2974613308906555
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,1,0.3299786647160848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,1,0.2979360024134318
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,1,0.3043253421783447
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,1,0.3032426635424296
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,1,0.3030400077501933
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,1,0.3338720003763835
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,1,0.32715733846028644
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,1,0.32679466406504315
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,1,0.32471466064453125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,1,0.3571253220240275
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,1,0.327349325021108
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,1,0.30102399984995526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,1,0.31909332672754925
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,1,0.33054933945337933
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,1,0.3268959919611613
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,1,0.32363200187683105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,1,0.3285333315531413
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,1,0.3283626635869344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,1,0.3351733287175496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,1,0.2989013393719991
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,1,0.30613332986831665
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,1,0.33222933610280353
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,1,0.330677330493927
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,1,0.35360534985860187
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,1,0.32607465982437134
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,1,0.32792532444000244
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,1,0.3347626527150472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,1,0.32127465804417926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,1,0.30131200949350995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,1,0.31868799527486164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,1,0.3009120027224223
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,1,0.34461331367492676
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,1,0.30003732442855835
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,1,0.35115734736124676
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,1,0.2970400055249532
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,1,0.33506667613983154
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,1,0.2974453369776408
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,1,0.3053013285001119
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,1,0.33763734499613446
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,1,0.30792532364527386
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,1,0.3280906677246094
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,1,0.3051946759223938
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,1,0.32445865869522095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,1,0.301530659198761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,1,0.36979198455810547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,1,0.30453334252039593
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,1,0.3285706639289856
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,1,0.32418133815129596
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,1,0.32922667264938354
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,1,0.3059413234392802
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,1,0.3280533353487651
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,1,0.3003679911295573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,1,0.572650671005249
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,1,0.31364800532658893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,1,0.33268799384435016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,1,0.30801600217819214
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,1,0.3503199815750122
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,1,0.3293173313140869
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,1,0.30431467294692993
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,1,0.3222133318583171
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,1,0.3224426706631978
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,1,0.3109546701113383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,1,1.0742186705271404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,1,0.5594240029652914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,1,0.3439733187357585
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,1,0.33507200082143146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,1,0.3278613289197286
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,1,0.3094399968783061
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,1,0.3160426616668701
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,1,0.3283519943555196
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,1,0.3086880048116048
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,1,0.32870932420094806
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,1,0.3181813359260559
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,1,0.3014240066210429
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,1,1.05404798189799
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,1,0.5574560165405273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,1,0.33894399801890057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,1,2.070085366566976
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,3,0.33773334821065265
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,3,0.35833601156870526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,3,0.3457546631495158
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,3,0.3356800079345703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,3,0.31016000111897785
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,1,0.7646293640136719
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,3,0.32659733295440674
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,3,0.3746933142344157
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,3,0.32073066631952923
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,3,0.4978400071461995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,3,0.3115786711374919
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,3,0.3476266860961914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,3,0.3375733296076457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,3,0.3469173510869344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,3,0.3468960126241048
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,1,0.3537280162175496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,1,0.35996798674265545
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,3,0.3428746859232585
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,3,0.3285760084788005
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,3,0.3038133382797241
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,1,0.30799466371536255
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,3,0.3034133315086365
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,3,0.3160746693611145
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,3,0.346288005510966
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,3,0.463098684946696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,3,0.37160531679789227
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,3,0.3666293223698934
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,3,0.3293439944585164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,3,0.3956799904505412
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,3,0.750378688176473
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,3,0.34143467744191486
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,3,1.0803146362304688
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,3,0.3489546775817871
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,3,0.3373386859893799
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,3,0.34938132762908936
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,3,0.3412693341573079
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,3,0.41948266824086505
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,3,0.3295626640319824
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,3,0.3418773412704468
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,3,0.3381013472874959
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,3,0.40215468406677246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,3,2.4678239822387695
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,3,0.37700267632802326
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,3,0.3415733178456624
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,3,0.3361599842707316
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,3,0.34300800164540607
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,3,0.33286933104197186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,3,0.30858665704727173
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,3,0.3240639964739482
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,3,0.9211946328481039
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,3,1.4766507148742676
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,3,0.3472586472829183
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,3,0.338917334874471
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,3,0.3325013319651286
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,3,0.34117865562438965
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,3,0.3457546631495158
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,3,0.34460266431172687
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,3,0.34888001283009845
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,3,0.6919786930084229
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,3,0.32894400755564374
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,3,0.3299786647160848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,3,0.3392373323440552
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,3,0.4729866584142049
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,3,0.5095413525899252
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,3,0.47208531697591144
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,3,0.33801066875457764
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,3,0.3113066752751668
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,3,0.3100373347600301
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,3,0.34330133597056073
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,3,0.3388693332672119
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,3,0.4714239835739136
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,3,0.33290133873621625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,3,3.4290078481038413
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,3,0.35567466417948407
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,3,0.3102293411890666
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,3,0.3075946569442749
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,3,0.47386666138966876
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,3,0.3322400053342183
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,3,0.34357865651448566
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,3,0.34115731716156006
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,3,0.3922400077184041
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,3,0.37411733468373615
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,3,0.9266186555226644
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,3,1.6984532674153645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,3,1.6879253387451172
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,3,0.3551679849624634
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,3,0.46987199783325195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,3,3.350096066792806
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,3,0.35816001892089844
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,3,0.3455679814020793
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,3,0.3450666666030884
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,3,0.8700746695200602
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,7,0.34326398372650146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,7,0.8807626565297445
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,7,0.5819253524144491
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,7,0.3448479970296224
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,7,0.35235734780629474
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,7,0.3412746588389079
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,7,0.36397333939870197
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,7,0.32577067613601685
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,7,0.34906665484110516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,7,0.33954131603240967
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,7,0.339466651280721
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,7,0.34089068571726483
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,7,0.33852267265319824
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,7,0.32594666878382367
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,7,0.33771200974782306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,7,0.33765868345896405
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,7,1.6203999519348145
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,7,0.5507306655248007
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,7,0.3370826641718547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,7,0.33929065863291424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,7,0.3258986671765645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,7,0.34109334150950116
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,7,0.30845866600672406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,7,0.3375093142191569
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,7,0.3469760020573934
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,7,0.33369068304697674
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,7,0.3102506597836812
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,7,0.3055359919865926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,7,0.30954132477442425
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,7,0.334112008412679
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,7,0.3066026568412781
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,7,0.3405119975407918
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,7,0.3311306635538737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,7,0.33990931510925293
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,7,0.3331093390782674
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,7,0.3560853401819865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,7,1.3599839210510254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,7,0.34651732444763184
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,7,0.7480853398640951
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,7,1.36845858891805
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,7,0.3408426841100057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,7,0.3460799853006999
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,7,0.40214399496714276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,7,0.3337013324101766
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,7,0.30193066596984863
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,7,0.3408213456471761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,7,0.3400799830754598
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,7,0.3314080039660136
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,7,0.35011200110117596
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,7,0.3356746832529704
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,7,0.3484853506088257
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,7,1.374783992767334
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,7,0.33577601114908856
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,7,0.3348906834920247
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,7,0.43430932362874347
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,7,0.3367893298467
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,7,0.3449866771697998
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,7,0.5594826539357504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,7,0.33532798290252686
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,7,0.33986131350199383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,7,0.33054933945337933
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,7,0.32922132809956867
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,7,0.34523733456929523
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,7,0.3460533221562703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,7,1.033573309580485
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,7,0.7864212989807129
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,7,0.329039990901947
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,7,0.33501867453257245
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,7,0.3372960090637207
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,7,0.3497920036315918
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,7,0.33985066413879395
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,7,0.5506879885991415
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,7,2.0223946571350098
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,7,1.6397919654846191
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,7,0.40379734834035236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,7,0.3426346778869629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,7,0.5603893200556437
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,7,0.3438719908396403
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,7,0.32224533955256146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,7,1.0423359870910645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,7,0.9715466499328613
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,7,4.016549428304036
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,7,0.4168533484141032
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,7,0.31795734167099
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,7,0.5554346640904745
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,7,1.0411146481831868
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,7,2.218575954437256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,7,0.4243786732355754
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,15,0.3336586554845174
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,15,0.33777066071828205
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,15,0.42747732003529865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,15,0.30266133944193524
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,15,0.357968012491862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,15,0.3428800106048584
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,15,1.510719935099284
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,15,1.692805290222168
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,15,0.32808534304300946
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,15,0.36295998096466064
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,15,0.3726666768391927
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,15,0.3603200117746989
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,15,0.3279306689898173
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,15,0.3397013346354167
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,15,0.3276960055033366
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,15,0.3445013364156087
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,15,0.33818666140238446
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,15,0.3301706711451213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,15,0.3396213452021281
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,15,0.34616533915201825
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,15,0.3295946717262268
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,15,0.3357439835866292
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,15,0.34122133255004883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,15,0.3381280104319255
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,15,0.33583998680114746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,15,0.34353601932525635
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,15,0.342901349067688
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,15,0.33897598584493
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,15,0.3699733416239421
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,15,0.3491733471552531
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,15,0.3365600109100342
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,15,0.33537065982818604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,15,0.3386720021565755
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,15,1.2039626439412434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,15,0.36994131406148273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,15,1.2350186506907146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,15,0.33511467774709064
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,15,0.3364959955215454
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,15,0.34538666407267254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,15,0.3657120068868001
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,15,0.3282986680666606
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,15,0.3352106809616089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,15,0.5836053291956583
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,15,0.3399999936421712
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,15,0.34348801771799725
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,15,0.3386559883753459
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,15,0.33665064970652264
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,15,0.30737600723902386
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,15,0.341050664583842
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,15,1.3564160664876301
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,15,0.3571733236312866
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,15,0.3447786569595337
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,15,0.3281973401705424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,15,1.1089599927266438
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,15,0.3455520073572795
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,15,0.3097440004348755
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,15,0.5598186651865641
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,15,0.3336000045140584
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,15,0.35609598954518634
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,15,0.3693386713663737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,15,0.33196266492207843
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,15,2.8056160608927407
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,15,0.36205331484476727
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,15,0.34721601009368896
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,15,0.33745066324869794
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,15,1.1147840023040771
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,15,1.3313013712565105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,15,0.6884053548177084
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,15,0.3831626574198405
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,15,0.3731253147125244
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,15,0.5836160182952881
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,15,0.3202880024909973
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,15,0.34279465675354004
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,15,0.5956053336461385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,15,1.054538647333781
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,15,0.343173344930013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,15,0.3427040179570516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,15,1.3458293279012044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,15,0.33739201227823895
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,15,2.0515467325846353
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,15,0.35629868507385254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,15,0.3474400043487549
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,15,0.4355146487553914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,15,2.037930647532145
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,15,1.0552586714426677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,15,0.42049598693847656
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,15,0.5550666650136312
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,15,4.058650652567546
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,31,0.33393065134684247
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,31,0.3590826590855916
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,31,0.33053867022196454
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,31,0.3468053340911865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,31,0.3325013319651286
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,31,0.33212800820668537
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,31,0.3452959855397542
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,31,0.34035201867421466
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,31,2.3219946225484214
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,31,0.3325600028038025
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,31,1.3303893407185872
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,31,1.9477599461873372
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,31,0.30160532395044964
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,31,0.33366934458414715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,31,0.30642666419347125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,31,0.37643734614054364
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,31,0.33404266834259033
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,31,0.3386346499125163
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,31,0.33400531609853107
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,31,0.3389386733373006
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,31,0.33849600950876874
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,31,0.3402880032857259
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,31,0.3516586621602376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,31,0.3711306651433309
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,31,0.33872532844543457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,31,0.33105067412058514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,31,0.33241067330042523
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,31,0.3336426814397176
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,31,1.354325294494629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,31,0.5839733282725016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,31,0.31147199869155884
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,31,0.5183093150456747
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,31,0.3359733422597249
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,31,0.4399679899215698
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,31,0.3288106719652812
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,31,1.3003626664479573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,31,0.33267199993133545
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,31,0.3064853350321452
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,31,0.3330079913139343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,31,0.8142879803975424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,31,0.3264426589012146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,31,2.098442713419596
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,31,0.3379093408584595
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,31,0.34122665723164874
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,31,0.3341013193130493
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,31,0.33166933059692383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,31,1.376021385192871
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,31,0.29702399174372357
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,31,0.35180266698201496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,31,0.3328160047531128
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,31,0.33274134000142414
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,31,0.33458133538564044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,31,0.3325013319651286
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,31,0.3192906578381856
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,31,0.3071039915084839
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,31,0.3323040008544922
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,31,0.34676265716552734
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,31,0.3558666706085205
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,31,0.34013867378234863
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,31,0.35020267963409424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,31,0.33716801802317303
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,31,0.5659893353780111
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,31,0.3041440049807231
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,31,0.3055359919865926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,31,0.33960533142089844
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,31,0.5639839967091879
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,31,0.33348266283671063
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,31,0.3272053400675456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,31,0.31893332799275714
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,31,0.3057013352711995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,31,1.0544533729553223
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,31,0.3268853425979614
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,31,0.3370186487833659
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,31,0.34414398670196533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,31,0.5701280037562052
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,31,2.0772053400675454
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,31,0.34809064865112305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,31,0.30371199051539105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,31,1.0661866664886475
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,31,0.33773334821065265
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,31,0.4231199820836385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,31,0.5650293429692587
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,31,0.35606932640075684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,31,1.0629119873046875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,31,0.34863468011220294
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,31,0.47335998217264813
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,31,4.083690643310547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,31,2.0515999794006348
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,63,0.3391786813735962
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,63,0.34065600236256915
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,63,0.3354239861170451
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,63,0.3344159921010335
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,63,0.3402880032857259
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,63,1.0897066593170166
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,63,0.3068319956461589
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,63,0.3287786642710368
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,63,0.36156264940897626
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,63,0.3265013297398885
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,63,1.0035733381907146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,63,0.3322346607844035
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,63,0.3338293234507243
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,63,0.34230931599934894
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,63,0.3269279996554057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,63,0.3277493317921956
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,63,0.3387519915898641
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,63,0.3297013243039449
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,63,1.0270559787750244
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,63,0.32578667004903156
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,63,0.33636267979939777
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,63,0.6976959705352783
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,63,0.3344800074895223
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,63,0.35312533378601074
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,63,0.334879994392395
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,63,0.33131200075149536
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,63,0.3343520164489746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,63,0.33877333005269367
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,63,0.3355040152867635
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,63,0.36760000387827557
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,63,0.3342133363087972
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,63,0.3375733296076457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,63,0.3386720021565755
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,63,0.32948267459869385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,63,0.3320373296737671
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,63,2.418922742207845
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,63,0.3421599864959717
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,63,0.3375626802444458
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,63,0.6608213186264038
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,63,0.3303999900817871
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,63,0.32894933223724365
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,63,0.3298719922701518
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,63,0.33182400465011597
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,63,0.6043733358383179
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,63,0.3776426712671916
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,63,0.3272266586621602
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,63,0.3395093282063802
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,63,0.47569068272908527
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,63,0.3608693281809489
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,63,1.3414239883422852
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,63,0.34377066294352215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,63,0.33794132868448895
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,63,0.34329601128896076
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,63,0.33555734157562256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,63,0.33907731374104816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,63,0.3052053252855937
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,63,0.4630933205286662
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,63,0.5635733207066854
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,63,0.333568016688029
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,63,0.3291200002034505
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,63,0.32868266105651855
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,63,0.32682667175928753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,63,0.33538134892781574
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,63,0.3415040175120036
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,63,0.33796266714731854
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,63,0.3237226605415344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,63,0.5111626784006754
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,63,0.5687786738077799
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,63,1.0925546487172444
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,63,0.3144746621449788
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,63,0.49861331780751544
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,63,0.29755733410517377
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,63,0.5713226795196533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,63,0.35469865798950195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,63,0.34031999111175537
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,63,1.0573973655700684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,63,2.071594715118408
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,63,0.3371573289235433
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,63,0.30456533034642536
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,63,0.32263465722401935
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,63,0.42657601833343506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,63,0.35525333881378174
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,63,1.0517813364664714
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,63,0.5618773301442465
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,63,4.063541412353516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,63,2.0862666765848794
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,63,0.35020267963409424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,63,0.5618559916814169
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,127,0.31725867589314777
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,127,0.30853333075841266
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,127,0.3169066707293193
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,127,0.43839999039967853
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,127,0.34780267874399823
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,127,0.34856534004211426
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,127,0.3455573320388794
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,127,0.30822400252024335
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,127,0.30245333909988403
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,127,0.3051626682281494
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,127,0.3327466646830241
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,127,0.34080533186594647
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,127,0.3391626675923665
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,127,0.3330826759338379
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,127,0.33138134082158405
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,127,0.33398934205373126
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,127,0.34889066219329834
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,127,0.3528053363164266
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,127,0.34226667881011963
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,127,0.3483039935429891
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,127,0.3349440097808838
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,127,0.3409759998321533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,127,0.33952001730600995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,127,0.33320534229278564
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,127,0.3436906735102336
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,127,0.3096853295962016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,127,0.3419413169225057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,127,0.3445546627044678
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,127,0.3179840048154195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,127,0.3256266713142395
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,127,0.3090026577313741
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,127,0.3396159807840983
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,127,0.3357386589050293
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,127,0.339408000310262
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,127,0.3057386676470439
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,127,0.33423467477162677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,127,0.30770667394002277
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,127,0.3093760013580322
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,127,0.33007999261220294
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,127,0.3314613302548726
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,127,0.30166399478912354
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,127,0.3019946614901225
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,127,0.3458026647567749
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,127,0.2982666691144307
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,127,0.33669865131378174
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,127,0.29967466990152997
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,127,0.3316799998283386
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,127,0.3044853409131368
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,127,0.30243732531865436
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,127,0.5723679860432943
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,127,0.3223680059115092
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,127,0.3067733248074849
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,127,0.3046293258666992
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,127,0.30726399024327594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,127,0.304421325524648
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,127,0.3040906588236491
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,127,0.2976906696955363
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,127,1.0805386702219646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,127,0.5738186836242676
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,127,0.33442668120066327
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,127,0.3418399890263875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,127,0.33978132406870526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,127,0.33560534318288165
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,127,0.33156800270080566
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,127,0.3311786651611328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,127,0.3083733320236206
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,127,2.1540533701578775
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,127,0.3073333303133647
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,127,0.3167840043703715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,127,0.43625601132710773
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,127,1.102346658706665
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,127,0.30955199400583905
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,127,0.5890239874521891
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,127,0.34122665723164874
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,127,0.32020799318949383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,127,0.35028799374898273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,127,0.3089226682980855
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,127,0.3038559953371684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,127,0.39498666922251385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,127,2.125216007232666
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,127,4.258042653401692
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,127,1.1023680369059246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,127,0.5794293483098348
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,255,0.3246346712112427
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,255,0.34754133224487305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,127,0.4513546625773112
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,255,0.32870932420094806
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,255,0.34886399904886883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,255,0.3375466664632161
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,127,0.37460800011952716
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,255,0.3393760124842326
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,255,0.31037867069244385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,255,0.3256640036900838
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,255,0.3325120011965434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,255,0.33371734619140625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,255,0.3344959815343221
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,255,0.3319466710090637
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,255,0.33085334300994873
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,255,0.33561599254608154
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,255,0.37513601779937744
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,127,0.36211733023325604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,255,0.3375893433888753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,127,0.37171733379364014
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,255,0.3758346637090047
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,255,0.33263999223709106
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,255,0.3376479943593343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,255,0.33928533395131427
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,255,0.3388799826304118
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,255,0.33099732796351117
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,255,0.33688533306121826
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,255,0.33322133620580036
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,255,0.34672534465789795
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,255,0.33816532293955487
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,255,0.33881600697835285
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,255,1.3713332811991374
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,255,0.3387093146642049
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,255,0.3044640024503072
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,255,0.34257598718007404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,255,0.3436373472213745
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,255,0.33243733644485474
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,255,0.3412533203760783
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,255,0.33431466420491535
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,255,1.344901402791341
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,255,0.33561599254608154
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,255,0.33817601203918457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,255,0.3362666765848796
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,127,0.7150719960530599
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,255,0.3351786533991496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,255,0.33905065059661865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,255,0.3320053418477376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,255,0.6551200151443481
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,255,0.3347626527150472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,255,0.9151679674784342
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,255,0.34303998947143555
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,255,0.330186665058136
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,255,0.3428320089975993
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,255,0.33345599969228107
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,255,0.3444853226343791
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,255,0.33752532800038654
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,255,0.39792001247406006
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,255,2.040719985961914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,255,0.3403786818186442
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,255,0.33633601665496826
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,255,0.3365600109100342
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,255,0.39134931564331055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,255,1.3736480077107747
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,255,0.3349013328552246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,255,0.7067093054453532
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,255,0.9413706461588541
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,255,0.33771733442942303
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,255,0.3345706860224406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,255,0.309605340162913
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,255,0.39528000354766846
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,255,0.7497440179189047
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,255,1.6751999855041504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,255,0.351034681002299
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,255,2.3653759956359863
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,255,0.3274880051612854
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,255,0.3394506772359212
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,255,1.3743146260579426
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,255,0.35552000999450684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,255,1.3644213676452637
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,255,0.4817013343175252
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,255,0.39690132935841876
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,255,0.33875731627146405
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,255,2.3784213066101074
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,255,0.3489813407262166
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,255,2.726319948832194
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,255,0.50164266427358
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,255,1.3830506006876628
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,255,0.47012801965077716
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,255,1.7162399291992188
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,255,0.7204000155131022
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,255,0.5873173475265503
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,511,0.3252906600634257
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,511,0.38027199109395343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,255,2.694234530131022
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,511,0.2956693371136983
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,511,0.3073386748631795
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,255,5.393893559773763
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,511,0.3352213303248088
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,511,0.3461013237635295
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,511,0.31726932525634766
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,511,1.5836639404296875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,511,0.3312480052312215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,511,0.3307093381881714
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,511,1.3540959358215332
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,511,2.3624000549316406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,511,1.9000959396362305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,511,0.3397866487503052
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,511,0.33905065059661865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,511,0.34138667583465576
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,511,0.3354560136795044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,511,1.2442879676818848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,511,0.32596800724665326
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,511,0.3457119862238566
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,511,0.3338293234507243
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,511,0.3097226619720459
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,511,0.3294293284416199
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,511,0.34222400188446045
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,511,0.34066665172576904
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,511,0.3343199888865153
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,511,0.3392159938812256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,511,0.32742400964101154
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,511,0.4041706720987956
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,511,0.3389600118001302
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,511,0.3404746850331624
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,511,0.3431946833928426
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,511,0.33424532413482666
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,511,2.251530647277832
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,511,0.33975998560587567
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,511,0.32500267028808594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,511,0.3807786703109741
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,511,0.33638401826222736
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,511,0.3410026629765828
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,511,0.374565323193868
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,511,0.3491893212000529
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,511,0.3314080039660136
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,511,1.3238986333211262
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,511,0.34221335252126056
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,511,0.3307200074195862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,511,0.33907731374104816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,511,0.337775985399882
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,511,0.3023573358853658
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,511,0.34095998605092365
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,511,0.35142401854197186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,511,0.3300693432490031
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,511,0.33642133076985675
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,511,0.3372480074564616
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,511,0.3761386473973592
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,511,1.0065600077311199
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,511,2.4346240361531577
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,511,0.560698668162028
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,511,2.4894933700561523
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,511,0.35354665915171307
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,511,0.32868266105651855
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,511,1.3506080309549968
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,511,0.33299734195073444
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,511,0.3153120080629985
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,511,0.9857813517252604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,511,0.5437119801839193
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,511,1.0565439860026042
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,511,1.5631786982218425
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,511,1.967184066772461
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,511,0.36451733112335205
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,511,0.3580586512883504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,511,0.44598933060963947
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,511,0.4790240128835042
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,511,0.5438400109608968
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,511,1.017621358235677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,511,0.8750560283660889
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,511,3.490997314453125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,511,0.4091626803080241
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,511,0.46916266282399494
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,511,0.424944003423055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,511,3.847536087036133
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,511,0.8520159721374512
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,511,0.7783786455790201
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,511,1.644015947977702
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,511,0.7347946961720785
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,511,1.9575893084208171
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,511,1.6413547197977703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,511,3.8367252349853516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,511,7.660677591959636
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,1023,0.30884265899658203
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,1023,0.3141653339068095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,1023,0.33694398403167725
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,1023,0.33613868554433185
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,1023,0.3047306736310323
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,1023,0.3148159980773926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,1023,0.349834680557251
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,1023,0.3221919933954875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,1023,0.3383786678314209
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,1023,0.3404746850331624
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,1023,0.3251039981842041
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,1023,0.3085813323656718
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,1023,0.3418080012003581
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,1023,0.3427199920018514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,1023,0.3020000060399373
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,1023,0.308186670144399
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,1023,0.49635199705759686
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,1023,0.338922659556071
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,1023,0.3167733351389567
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,1023,0.38523733615875244
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,1023,0.34200533231099445
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,1023,0.31253333886464435
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,1023,0.3335786660512288
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,1023,0.34116800626118976
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,1023,0.3444746732711792
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,1023,0.3142079909642537
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,1023,0.3447253306706746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,1023,0.33211733897527057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,1023,0.3076373338699341
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,1023,0.3249280055363973
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,1023,0.34781865278879803
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,1023,0.33344535032908124
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,1023,0.34384000301361084
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,1023,0.33421866099039715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,1023,0.4498506784439087
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,1023,0.3321546713511149
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,1023,0.33109333117802936
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,1023,0.30528000990549725
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,1023,0.3102826674779256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,1023,0.30963200330734253
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,1023,0.306768000125885
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,1023,0.792954683303833
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,1023,0.4471733172734578
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,1023,0.3728319803873698
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,1023,0.32815466324488324
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,1023,0.33126399914423627
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,1023,0.33289599418640137
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,1023,0.3354719877243042
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,1023,0.3142613371213277
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,1023,1.5929066340128581
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,1023,0.8342133363087972
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,1023,0.46963731447855633
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,1023,0.3519786596298218
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,1023,0.34459201494852704
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,1023,0.3465813398361206
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,1023,0.3458346525828044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,1023,0.29527999957402545
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,1023,3.1085707346598306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,1023,1.5938560167948406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,1023,0.8277066548665365
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,1023,0.4591253201166789
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,1023,0.41198933124542236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,1023,0.38973867893218994
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,1023,0.3808799982070923
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,1023,0.8686559995015463
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,1023,0.31364800532658893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,1023,0.3130613366762797
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,1023,6.124575932820638
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,1023,0.314245343208313
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,1023,3.0995572408040366
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,1023,0.3157920042673747
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,1023,0.834496021270752
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,1023,1.5936800638834636
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,1023,0.7380053202311198
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,1023,0.3162720004717509
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,1023,0.6959359645843506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,1023,0.677285353342692
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,1023,1.6494879722595215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,1023,0.3611253499984741
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,1023,6.106128056844075
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,1023,12.19716771443685
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,1023,3.096538543701172
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,1023,1.3820160230000813
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,2047,0.36722131570180255
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,1023,1.305578629175822
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,2047,0.3468159834543864
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,2047,0.3361599842707316
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,2047,0.333349347114563
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,2047,0.3431253433227539
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,2047,0.3507466713587443
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,2047,0.37774932384490967
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,2047,0.33855998516082764
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,2047,0.34223465124766034
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,2047,0.36690668265024823
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,2047,0.30425600210825604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,2047,0.34673066933949787
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,2047,0.35439467430114746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,2047,0.3783200184504191
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,2047,0.37836798032124835
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,1023,1.5940799713134766
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,2047,0.3564213514328003
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,2047,0.3267520070075989
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,2047,0.309279998143514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,2047,0.34355731805165607
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,2047,0.31642667452494305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,2047,0.3452426592508952
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,2047,0.34510934352874756
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,2047,0.35358933607737225
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,2047,0.341536005338033
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,2047,0.3447146813074748
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,2047,0.36150399843851727
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,2047,0.3434293270111084
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,2047,0.526528000831604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,2047,0.33499733606974286
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,1023,1.267807960510254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,2047,0.33131200075149536
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,2047,0.7796693642934164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,2047,0.3567626476287842
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,2047,0.3335626522699992
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,2047,0.33483731746673584
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,2047,0.3494986693064372
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,2047,0.41975998878479004
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,2047,0.3402666648228963
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,1023,3.2461067835489907
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,2047,0.30797866980234784
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,2047,0.3407466808954875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,2047,0.3652373154958089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,2047,0.4084746837615967
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,2047,0.7400533358256022
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,2047,0.6973013083140055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,2047,0.7404692967732748
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,2047,0.3355199893315633
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,2047,0.34622931480407715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,2047,0.3065653244654338
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,2047,0.36745599905649823
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,2047,0.9567519823710123
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,2047,0.4188266595204671
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,2047,0.3527626593907674
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,2047,0.32954132556915283
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,2047,1.358239968617757
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,2047,0.3229973316192627
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,2047,0.35733334223429364
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,2047,0.6615253289540609
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,1023,0.5374933481216431
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,2047,0.7717119852701823
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,2047,1.4264960289001465
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,2047,0.42418134212493896
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,2047,0.37990399201711017
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,2047,0.3729066848754883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,2047,2.744175910949707
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,2047,0.37300801277160645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,2047,0.9973386923472086
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,2047,0.7625973224639893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,2047,1.4075627326965332
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,2047,2.738704045613607
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,2047,0.6969546476999918
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,2047,0.6759626865386963
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,2047,0.6658399899800619
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,2047,5.406063715616862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,2047,1.6503626505533855
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,2047,1.416752020517985
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,2047,2.7401599884033203
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,2047,1.2719573179880779
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,2047,1.3124000231424968
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,2047,1.217845360438029
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,2047,5.376170476277669
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,2047,10.66375478108724
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,2047,3.205850601196289
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,2047,2.755685488382975
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,2047,2.4586240450541177
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,4095,0.343178669611613
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,2047,2.3735040028889975
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,2047,5.386751810709636
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,4095,0.33907731374104816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,4095,0.34060267607371014
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,4095,0.34141866366068524
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,2047,2.3384639422098794
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,4095,0.3398880163828532
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,4095,0.33056533336639404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,4095,0.32453866799672443
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,4095,0.9890186786651611
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,4095,0.345909317334493
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,4095,0.3426080147425334
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,4095,0.3410133520762126
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,4095,0.33617599805196124
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,4095,0.3447466691335042
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,4095,0.33610133330027264
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,4095,0.3338079849878947
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,4095,0.9047573407491049
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,2047,21.251995086669922
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,4095,0.34705066680908203
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,4095,0.34511999289194745
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,4095,0.34337600072224933
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,2047,6.263050715128581
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,4095,0.3942879835764567
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,4095,0.38016001383463544
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,4095,0.37882133324941
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,4095,1.2548106511433919
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,4095,0.45254401365915936
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,4095,0.3658133347829183
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,4095,0.3768959840138753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,4095,0.3388693332672119
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,4095,0.34570133686065674
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,4095,0.3478879928588867
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,4095,0.7998186747233073
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,4095,0.3512959877649943
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,4095,0.35120534896850586
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,4095,0.34193066755930585
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,4095,1.4399360020955403
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,4095,0.3078666726748149
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,4095,0.35678398609161377
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,4095,0.7073386510213217
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,4095,0.3439573446909587
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,4095,0.3821920156478882
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,4095,0.9226933320363363
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,4095,0.44676800568898517
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,4095,0.4933866659800212
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,4095,1.3253920078277588
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,4095,0.3678933382034302
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,4095,0.36745067437489826
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,4095,0.6982186635335287
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,4095,0.3612639904022217
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,2047,10.658143997192383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,4095,0.6628746589024862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,4095,0.7129920323689779
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,4095,0.45774932702382404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,4095,1.3277386824289958
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,4095,0.42139732837677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,4095,0.4160533348719279
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,4095,2.4698774019877114
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,4095,0.41102401415507
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,4095,1.2553386688232422
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,4095,1.372538725535075
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,4095,0.7326506773630778
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,4095,0.669648011525472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,4095,0.6556906700134277
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,4095,2.5638720194498696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,4095,0.6607253154118856
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,4095,5.082111994425456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,4095,1.8986399968465169
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,4095,1.2607253392537434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,4095,1.3809706370035808
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,4095,2.5583839416503906
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,4095,1.2310506502787273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,4095,5.055909474690755
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,4095,1.228495995203654
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,4095,9.974159876505533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,4095,2.581029256184896
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,4095,3.2061440149943032
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,4095,2.3620427449544272
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,4095,5.122384071350098
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,4095,2.3136159578959146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,4095,9.973189036051432
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,4095,19.82733917236328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,4095,6.315653483072917
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,4095,2.292725404103597
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1024,1,8,4095,5.088122685750325
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1024,1,16,4095,4.6044158935546875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,8191,0.32758933305740356
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,8191,0.35301868120829266
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,8191,0.3573013146718343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1024,1,4,4095,9.970698674519857
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,8191,0.327290674050649
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,8191,0.3596213261286418
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,8191,0.3501013517379761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,8191,0.6483946641286215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,8191,0.36338667074839276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,8191,0.3644160032272339
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1024,1,32,4095,4.525600115458171
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1024,1,2,4095,19.87722651163737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,8191,0.35951467355092365
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,8191,0.36087465286254883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,8191,0.35262401898701984
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,8191,0.36566932996114093
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,8191,0.3537760178248088
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,8191,0.32576533158620197
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,8191,0.356112003326416
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,8191,0.3680746555328369
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,8191,0.36925868193308514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,8191,1.399354616800944
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,8191,0.552570660909017
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1024,1,64,4095,4.4917707443237305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,8191,0.3901919921239217
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,8191,0.38815466562906903
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,8191,0.3492106596628825
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,8191,1.0944693088531494
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,8191,0.7935946782430013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,8191,0.5519520044326782
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,8191,0.3602026700973511
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,8191,0.3593706687291463
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,8191,0.3668053150177002
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,8191,0.3566879828770955
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,8191,0.3529706796010335
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1024,1,1,4095,39.474507649739586
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,8191,1.105242649714152
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,8191,0.353279987970988
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,8191,0.5499306519826254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,8191,0.3377866744995117
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,8191,0.8025493621826172
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,8191,1.2680959701538086
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,8191,0.35333867867787677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1024,1,128,4095,12.302555084228516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,8191,0.3623146613438924
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,8191,1.1148746808369954
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,8191,1.2777173519134521
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,8191,0.7952426274617513
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,8191,2.4741172790527344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,8191,2.4022666613260903
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,8191,0.514794667561849
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,8191,0.5173919995625814
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,8191,0.5244693358739217
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,8191,1.256826639175415
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,8191,2.49563201268514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,8191,1.3111626307169597
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,8191,0.7557066281636556
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,8191,0.7490879694620768
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,8191,4.676464080810547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,8191,0.8206559816996256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,8191,0.7557760079701742
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,8191,2.4309813181559243
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,8191,1.3437919616699219
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,8191,1.2189813454945881
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,8191,1.2385546366373699
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,8191,2.583258628845215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,8191,1.2254026730855305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,8191,4.885093371073405
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,8191,3.708863894144694
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,8191,9.652725219726562
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,8191,19.161253611246746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,8191,2.5942187309265137
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,8191,2.3563040097554526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,8191,2.4054719607035318
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,8191,2.3867732683817544
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,8191,6.331322352091472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,8191,4.860538800557454
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,8191,9.640538533528646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,512,1,16,8191,4.519941329956055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,512,1,8,8191,4.931685447692871
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,512,1,4,8191,9.642762502034506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,512,1,64,8191,4.443327903747559
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,8191,4.481274604797363
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,512,1,2,8191,19.133738199869793
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,512,1,128,8191,12.54642105102539
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,512,1,1,8191,38.15893300374349
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,16383,0.5425546566645304
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,16383,0.4774880011876424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,16383,0.4294240077336629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,16383,0.42106131712595624
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,16383,0.42056000232696533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,16383,1.2385706901550293
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,16383,0.5503573417663574
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,16383,0.5331573486328125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,16383,0.5303359826405843
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,16383,0.4702719847361247
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,16383,0.46671466032663983
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,16383,0.46954667568206787
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,16383,2.0131680170694985
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,16383,1.0014346440633137
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,16383,0.5429226557413737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,16383,0.5366719961166382
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,16383,0.4751733144124349
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,16383,0.47179198265075684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,16383,0.47129066785176593
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,16383,2.085653305053711
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,16383,1.466383934020996
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,16383,0.5559146801630656
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,16383,0.4811520179112752
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,16383,0.479039986928304
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,16383,2.132314682006836
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,16383,2.4058772722880044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,16383,1.4761013984680176
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,16383,1.0095307032267253
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,16383,0.554805318514506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,16383,0.5053866704305013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,16383,0.5024480024973551
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,16383,0.5017439921696981
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,16383,2.1559786796569824
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,16383,4.758624076843262
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,16383,2.425312042236328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,16383,1.4876267115275066
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,16383,1.0420107046763103
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,16383,0.9563146432240804
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,16383,0.9596052964528402
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,16383,0.959882656733195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,16383,2.4105493227640786
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,16383,9.113791783650717
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,16383,0.5338240067164103
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,16383,4.822928110758464
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,16383,1.5496692657470703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,16383,2.4999094009399414
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,16383,1.4356907208760579
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,16383,1.4334400494893391
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,16383,1.4258774121602376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,16383,4.7907358805338545
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,16383,9.528703689575195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,16383,18.917743682861328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,16383,2.565989335378011
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,16383,5.004426638285319
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,16383,2.3715786933898926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,16383,2.363093376159668
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,16383,2.350053310394287
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,16383,7.338101069132487
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,16383,0.5223093430201212
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,256,1,1,16383,37.535675048828125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,256,1,4,16383,9.454298655192057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,256,1,8,16383,5.030890782674153
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,256,1,16,16383,4.665050824483235
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,256,1,2,16383,18.94168472290039
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,256,1,32,16383,4.615797360738118
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,16383,0.5311040083567301
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,256,1,64,16383,4.637226740519206
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,256,1,128,16383,12.587776184082031
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,16383,0.5325599908828735
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,16383,0.5468106667200724
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,16383,0.4817173480987549
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,16383,1.00054931640625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,512,1,32,16383,8.81551996866862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,32767,0.9780159791310629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,32767,0.9845600128173828
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,32767,0.8571306864420573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,32767,0.9915359814961752
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,32767,0.855557362238566
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,32767,3.9093812306722007
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,32767,0.9862720171610514
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,32767,0.9813066323598226
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,32767,1.0720586776733398
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,32767,0.8658080101013184
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,32767,0.994767983754476
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,32767,1.99782927831014
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,32767,3.9630985260009766
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,32767,0.9928586483001709
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,32767,1.060762643814087
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,32767,0.8772213459014893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,32767,0.8783626556396484
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,32767,0.8726186752319336
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,32767,2.827365239461263
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,32767,4.073786735534668
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,32767,1.9483253161112468
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,32767,1.0320320129394531
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,32767,0.893999973932902
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,32767,1.0099146366119385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,32767,0.886853297551473
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,32767,0.8895626862843832
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,32767,4.17249075571696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,32767,4.6802825927734375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,32767,1.929957389831543
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,32767,1.0268212954203289
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,32767,2.855093320210775
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,32767,0.9347146352132162
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,32767,0.9253066380818685
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,32767,0.9302399953206381
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,32767,4.242885271708171
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,32767,4.7540585199991865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,32767,2.892554601033529
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,32767,9.342453638712565
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,32767,2.001941363016764
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,32767,1.8480265935262044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,32767,1.8526879946390789
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,32767,1.8334132830301921
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,32767,4.752170562744141
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,32767,4.876282691955566
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,32767,3.0072692235310874
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,32767,9.60644785563151
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,32767,2.7980213165283203
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,32767,2.85861873626709
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,32767,18.287743886311848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,32767,2.782069206237793
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,32767,9.491493225097656
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,128,1,4,32767,9.80888557434082
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,128,1,2,32767,18.720128377278645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,128,1,8,32767,5.003679911295573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,128,1,16,32767,4.657466570536296
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,128,1,1,32767,37.378011067708336
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,128,1,32,32767,4.624688148498535
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,128,1,64,32767,4.618032137552897
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,128,1,128,32767,14.631749471028646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,65535,1.8815946578979492
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,1,1,16,65535,1.66102933883667
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,1,1,8,65535,1.8889493942260742
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,1,1,32,65535,1.6517972946166992
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,1,1,64,65535,1.6496159235636394
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,65535,1.8810240427652996
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,1,1,128,65535,7.782837549845378
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,65535,1.8851253191630046
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,65535,1.6658506393432617
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,65535,1.9079039891560872
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,65535,1.6584854125976562
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,65535,1.65558926264445
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,65535,1.9052747090657551
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,65535,1.9142239888509114
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,65535,7.842250823974609
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,65535,1.6843199729919434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,65535,1.6681920687357585
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,65535,5.53386116027832
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,65535,8.081594467163086
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,65535,1.98743470509847
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,65535,3.720645268758138
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,65535,1.7029333114624023
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,65535,1.9437813758850098
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,65535,1.6971786816914876
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,65535,1.6954719225565593
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,65535,8.298095703125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,65535,9.236293156941732
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,65535,3.7768532435099282
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,65535,5.579077402750651
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,65535,1.7842186292012532
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,65535,1.9809120496114094
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,65535,1.7841973304748535
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,65535,1.7805919647216797
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,65535,8.422143936157227
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,65535,9.36126963297526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,65535,18.609269460042317
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,65535,3.6483519872029624
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,65535,3.5953919092814126
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,65535,3.6092052459716797
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,65535,9.405445098876953
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,64,1,2,65535,19.312816619873047
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,64,1,1,65535,36.84474182128906
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,64,1,4,65535,9.638938903808594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,64,1,8,65535,5.918570836385091
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,64,1,16,65535,5.538607915242513
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,64,1,32,65535,5.455600102742513
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,64,1,64,65535,5.477509180704753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,64,1,128,65535,18.900821685791016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,65535,1.8939520517985027
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,65535,1.9032799402872722
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,65535,1.670805295308431
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,65535,3.701471964518229
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,65535,4.017290751139323
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,65535,5.69108772277832
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,1,1,1,131071,3.6767520904541016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,2,1,1,131071,3.70032533009847
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,2,1,2,131071,3.6921653747558594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,2,1,4,131071,3.707333246866862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,2,1,8,131071,3.720970789591471
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,2,1,16,131071,3.261530558268229
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,1,1,4,131071,3.6774559020996094
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,2,1,32,131071,3.2460854848225913
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,1,1,2,131071,3.6949332555135093
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,2,1,64,131071,3.240629196166992
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,2,1,128,131071,15.680362701416016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,4,1,2,131071,3.7132533391316733
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,4,1,1,131071,7.3273970286051435
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,4,1,4,131071,3.736490567525228
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,4,1,8,131071,3.7582454681396484
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,4,1,16,131071,3.286106745402018
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,4,1,32,131071,3.2737013498942056
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,4,1,64,131071,3.2705599466959634
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,8,1,2,131071,7.36894416809082
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,8,1,1,131071,10.940053304036459
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,4,1,128,131071,16.181007385253906
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,8,1,4,131071,3.880730628967285
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,8,1,8,131071,3.8033278783162436
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,8,1,16,131071,3.343173344930013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,8,1,32,131071,3.3282718658447266
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,8,1,64,131071,3.312293370564779
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,8,1,128,131071,16.597279866536457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,16,1,2,131071,11.04415512084961
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,16,1,1,131071,18.37708791097005
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,16,1,8,131071,3.8651787439982095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,16,1,4,131071,7.480202356974284
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,16,1,16,131071,3.470047950744629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,16,1,32,131071,3.4827839533487954
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,16,1,64,131071,3.493626594543457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,16,1,128,131071,16.639760335286457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,64,32,1,2,131071,18.75401560465495
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,32,32,1,4,131071,11.238058725992838
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,8,32,1,16,131071,7.231237411499023
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,16,32,1,8,131071,7.762304306030273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,128,32,1,1,131071,37.48292795817057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,4,32,1,32,131071,7.184821446736653
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,2,32,1,64,131071,7.167061487833659
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,generation_mla,vllm_triton_mla,float16,float16,1,32,1,128,131071,18.70294443766276
